id
int64 393k
2.82B
| repo
stringclasses 68
values | title
stringlengths 1
936
| body
stringlengths 0
256k
⌀ | labels
stringlengths 2
508
| priority
stringclasses 3
values | severity
stringclasses 3
values |
|---|---|---|---|---|---|---|
97,431,757 | opencv | Training models for the Latent SVM detector | Transferred from http://code.opencv.org/issues/3658
```
|| David W. on 2014-04-21 07:58
|| Priority: Normal
|| Affected: branch '2.4' (2.4-dev)
|| Category: objdetect
|| Tracker: Bug
|| Difficulty: Easy
|| PR:
|| Platform: Any / Any
```
## Training models for the Latent SVM detector
```
Hi there! :)
The OpenCV implementation of the Latent SVM detector [1] from the objdetect module works great with the prebuild models from the opencv_extra repository [2].
But how can you train your own models? Some information on this topic in the OpenCV documentation would be very helpful. How were the models from opencv_extra generated?
Does somebody know the relationship between the OpenCV implementation and the "original" one [3]? Unfortunatly the author does not know [4].
I hope that we have somebody in the community who knows something about this stuff.
[1] http://docs.opencv.org/master/modules/objdetect/doc/latent_svm.html
[2] https://github.com/Itseez/opencv_extra/tree/master/testdata/cv/latentsvmdetector
[3] http://www.cs.berkeley.edu/~rbg/latent/
[4] https://github.com/rbgirshick/voc-dpm/issues/4
```
## History
##### Dmitry Retinskiy on 2014-09-17 10:16
```
Hi David,
I'd suggest asking your question at this forum http://answers.opencv.org.
- Affected version changed from 2.4.8 (latest release) to branch '2.4'
(2.4-dev)
```
##### Steven Puttemans on 2014-09-22 13:47
```
If you want a new model, you need to download the original software of Felzenswalb.
It has a training interface with it.
Those models can be used in OpenCV. There is no OpenCV dedicated training interface provided.
The models in the extra folder are just retrieved from the Pascal VOC Challenge.
```
| bug,auto-transferred,priority: normal,affected: 2.4,category: objdetect | low | Critical |
97,431,790 | opencv | openCV support for multiple camera instance in Android | Transferred from http://code.opencv.org/issues/3660
```
|| ganesh manvizhi on 2014-04-22 15:44
|| Priority: High
|| Affected: None
|| Category: android
|| Tracker: Feature
|| Difficulty: Hard
|| PR:
|| Platform: None / None
```
## openCV support for multiple camera instance in Android
```
Is it possible to access multiple camera (4 cameras in my case) in openCV on Android platform. If so, Please also mention the Android versions that supports this feature of openCV.
Please let me clarify this issue ASAP.
Regards,
Ganesh.M
```
## History
##### Alexander Smorkalov on 2014-04-30 19:05
```
- Target version changed from 2.4.9 to 2.4.10
```
##### Dmitry Retinskiy on 2014-05-29 07:23
```
Hi Ganesh,
I'd recommend to ask the community at Q&A forum:
http://answers.opencv.org/questions/
```
##### Ayberk Özgür on 2014-09-29 12:56
```
It should be possible in theory, as long as those 4 cameras are all connected to different USB buses (which is highly unlikely).
```
| auto-transferred,feature,priority: low,category: videoio(camera),platform: android | low | Minor |
97,431,845 | opencv | Failed to cross-compile with mingw-w64 | Transferred from http://code.opencv.org/issues/3661
```
|| Alexey Gornostaev on 2014-04-23 06:19
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: build/install
|| Tracker: Bug
|| Difficulty: Easy
|| PR:
|| Platform: x64 / Linux
```
## Failed to cross-compile with mingw-w64
```
https://github.com/Itseez/opencv/blob/master/modules/core/src/parallel.cpp
45: #if defined WIN32 || defined WINCE
should looks like
45: #if defined WIN32 || defined WINCE || defined _WIN32
cause SYSTEM_INFO structure is undefined here elsewhere:
454: #if defined WIN32 || defined _WIN32
455: SYSTEM_INFO sysinfo;
```
## History
##### Alexey Gornostaev on 2014-04-23 06:22
```
another one: highgui/src/cap_ffmpeg.cpp using HMODULE declaration but not includes windows-specific headers
```
##### Dmitry Retinskiy on 2014-05-30 05:27
```
Hi Roman,
could you check this problem?
Thanks.
- Status changed from New to Open
```
##### Philippe FOUBERT on 2014-05-30 10:13
```
Hi,
Just for information: I am also using mingw-w64 on branch 'master' (3.0-dev) (update at least once a week) and I don't have any compilation problem.
+Details:+
* mingw : gcc version 4.8.2, x86_64-win32-seh, Built by MinGW-W64 project
* OS : Windows 7 Professionnel, Service Pack 1, 64 bits
* branch : https://github.com/philippefoubert/opencv.git
* computer: Intel core i7-3740QM CPU @2.70GHz
Regards,
Philippe.
```
| bug,auto-transferred,priority: normal,category: build/install,affected: 3.4 | low | Critical |
97,431,889 | opencv | OpenCV Python unable to access correct OpenNI device channels | Transferred from http://code.opencv.org/issues/3663
```
|| Andrew Braun on 2014-04-24 03:48
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: highgui-camera
|| Tracker: Bug
|| Difficulty: Medium
|| PR:
|| Platform: x64 / Linux
```
## OpenCV Python unable to access correct OpenNI device channels
```
I put together some simple code in python to grab different channels from OpenNI devices. I built OpenCV myself with all the PrimeSense and OpenNI support enabled. The OpenNI samples work perfectly for both the Kinect sensor and PrimeSense sensor, as well as the OpenCV samples for testing OpenNI support (./cpp-example-openni_capture).
Here is the code I put together.
@
import cv2
import cv2.cv as cv
capture = cv2.VideoCapture(cv.CV_CAP_OPENNI)
capture.set(cv.CV_CAP_OPENNI_IMAGE_GENERATOR_OUTPUT_MODE, cv.CV_CAP_OPENNI_VGA_30HZ)
print capture.get(cv.CV_CAP_PROP_OPENNI_REGISTRATION)
while True:
if not capture.grab():
print "Unable to Grab Frames from camera"
break
okay1, depth_map = capture.retrieve(cv.CV_CAP_OPENNI_DEPTH_MAP)
if not okay1:
print "Unable to Retrieve Disparity Map from camera"
break
okay2, gray_image = capture.retrieve(cv.CV_CAP_OPENNI_GRAY_IMAGE)
if not okay2:
print "Unable to retrieve Gray Image from device"
break
cv2.imshow("depth camera", depth_map)
cv2.imshow("rgb camera", gray_image)
if cv2.waitKey(10) == 27:
break
cv2.destroyAllWindows()
capture.release()
@
So everything runs fine, but the results that are being displayed are not the correct channels... For example if I wanted to access the gray image channel and the depth map channel, both images being displayed are depth_maps.
Yes I've tried accessing other channels and changing the OPENNI_IMAGE_GENERATOR_MODE. Unfortunately the results have stayed consistent. No matter what I try I always get the same depth channel back. depth_map-gray_image yields an all black image.
Like I said the C++ OpenCV OpenNI examples all work perfectly for both the Kinect sensor and primesense sensor. It seems like a problem with the Python modules, or I am doing something really stupid.
Running on Ubuntu 12.04 LTS OpenCV version 2.4.9
Thanks for helping.
Drew
```
## History
##### Alexander Smorkalov on 2014-04-30 19:05
```
- Target version changed from 2.4.9 to 2.4.10
```
##### Dmitry Retinskiy on 2014-09-17 10:12
```
Alexander, could you check this?
Thanks.
- Affected version changed from 2.4.8 (latest release) to 2.4.9
(latest release)
- Assignee set to Alexander Smorkalov
- Status changed from New to Open
```
| bug,auto-transferred,priority: normal,affected: 2.4,category: videoio(camera) | low | Critical |
97,431,950 | opencv | python cv2.floodFill Segmentation Fault on large images | Transferred from http://code.opencv.org/issues/3669
```
|| Michael Goldman on 2014-04-30 16:07
|| Priority: Normal
|| Affected: 2.4.0 - 2.4.8
|| Category: imgproc, video
|| Tracker: Bug
|| Difficulty: Medium
|| PR:
|| Platform: x64 / Linux
```
## python cv2.floodFill Segmentation Fault on large images
```
Performing python floodFill on large images ~2Gigapixels yields segmentation fault. Have tried on Ubuntu 12.04 with 12GB memory and RedHat 5.10 with ~96GB memory.
-- success
% python floodfill_repro.py 32768 32768
-- success
% python floodfill_repro.py 46341 46341
-- fail
% python floodfill_repro.py 46341 46342
-- fail
% python floodfill_repro.py 65535 65535
```
## History
##### Alexander Shishkov on 2014-05-06 20:19
```
- Affected version changed from 2.4.8 (latest release) to 2.4.0 -
2.4.8
- Difficulty set to Medium
- Assignee set to Vadim Pisarevsky
- Category set to imgproc, video
```
##### Andrew Senin on 2014-05-07 15:49
```
- Status changed from New to Open
```
| bug,auto-transferred,priority: normal,category: imgproc,category: video,affected: 2.4 | low | Critical |
97,432,019 | opencv | Releasing VideoCapture / Webcam doesn't work | Transferred from http://code.opencv.org/issues/3672
```
|| Martijn van der Veen on 2014-05-02 21:02
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: python bindings
|| Tracker: Bug
|| Difficulty: Medium
|| PR:
|| Platform: x64 / Linux
```
## Releasing VideoCapture / Webcam doesn't work
```
Using python (3.4) bindings on Ubuntu 14.04 with build-in webcam using v4l4, master branch:
In [2]: cap = cv2.VideoCapture(0)
HIGHGUI ERROR: V4L/V4L2: VIDIOC_S_CROP
In [3]: a,b = cap.read() # although error, reading frame works (a=True, b=np.array)
In [4]: cap.release()
In [5]: cap.isOpened()
Out[5]: False
In [6]: cap = cv2.VideoCapture(0)
HIGHGUI ERROR: V4L2: Pixel format of incoming image is unsupported by OpenCV
In [7]: a,b = cap.read()
VIDIOC_STREAMON: Bad file descriptor
After calling .release(), webcam light stays on and webcam stays unreachable for other programs (e.g., Cheese).
```
## History
##### Martijn van der Veen on 2014-05-02 21:24
```
(Btw, the webcam _is_ released when python exits, but in some cases that's not enough - e.g., for running programs that need to release the webcam now and then.)
```
##### Alexander Smorkalov on 2014-05-04 15:58
```
- Category set to python bindings
```
##### Andrew Senin on 2014-05-06 14:21
```
Hello Martijn,
Since our resources are limited please consider investigating this issue and creating a fixing pull request (http://code.opencv.org/projects/opencv/wiki/How_to_contribute). All help to the project is highly appreciated!
- Difficulty set to Medium
- Assignee set to Martijn van der Veen
- Status changed from New to Open
```
##### Maksim Shabunin on 2015-04-27 09:12
```
- Target version changed from 3.0-beta to 3.0
```
##### Maksim Shabunin on 2015-04-28 13:42
```
Can not reproduce with the 3.0-rc1 release. Webcam is successfully released and reopened.
Can you, please, try to reproduce it with the latest version? Probably it has been fixed already.
- Assignee changed from Martijn van der Veen to Maksim Shabunin
```
##### Maksim Shabunin on 2015-04-28 13:42
```
- Status changed from Open to Incomplete
```
| bug,auto-transferred,priority: normal,category: python bindings,affected: 3.4 | low | Critical |
97,432,230 | opencv | GraphCutSeamFinderBase has no virtual destructor | Transferred from http://code.opencv.org/issues/3685
```
|| Alexey Gornostaev on 2014-05-12 04:55
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: stitching
|| Tracker: Bug
|| Difficulty: Easy
|| PR:
|| Platform: x64 / Linux
```
## GraphCutSeamFinderBase has no virtual destructor
```
Stitching module failed to build with -Werror=non-virtual-dtor
```
## History
##### Dmitry Retinskiy on 2014-05-29 06:23
```
Hi Leonid,
could you check this problem?
Thanks.
- Assignee changed from Roman Donchenko to Leonid Beynenson
- Category changed from build/install to stitching
```
##### Dmitry Retinskiy on 2014-05-30 07:42
```
- Status changed from New to Open
```
| bug,auto-transferred,priority: normal,affected: 3.4,category: stitching | low | Critical |
97,432,265 | opencv | opencv.pc links to wrong TBB | Transferred from http://code.opencv.org/issues/3691
```
|| Thorsten Ball on 2014-05-14 06:17
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: build/install
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Mac OSX
```
## opencv.pc links to wrong TBB
```
On OS X 10.9.2 the generated `opencv.pc` does not link to the correct TBB library.
@# Package Information for pkg-config
prefix=/usr/local
exec_prefix=${prefix}
libdir=
includedir_old=${prefix}/include/opencv
includedir_new=${prefix}/include
Name: OpenCV
Description: Open Source Computer Vision Library
Version: 2.4.9
Libs: ${exec_prefix}/lib/libopencv_calib3d.dylib ${exec_prefix}/lib/libopencv_contrib.dylib ${exec_prefix}/lib/libopencv_core.dylib ${exec_prefix}/lib/libopencv_features2d.dylib ${exec_prefix}/lib/libopencv_flann.dylib ${exec_prefix}/lib/libopencv_gpu.dylib ${exec_prefix}/lib/libopencv_highgui.dylib ${exec_prefix}/lib/libopencv_imgproc.dylib ${exec_prefix}/lib/libopencv_legacy.dylib ${exec_prefix}/lib/libopencv_ml.dylib ${exec_prefix}/lib/libopencv_nonfree.dylib ${exec_prefix}/lib/libopencv_objdetect.dylib ${exec_prefix}/lib/libopencv_ocl.dylib ${exec_prefix}/lib/libopencv_photo.dylib ${exec_prefix}/lib/libopencv_stitching.dylib ${exec_prefix}/lib/libopencv_superres.dylib ${exec_prefix}/lib/libopencv_ts.a ${exec_prefix}/lib/libopencv_video.dylib ${exec_prefix}/lib/libopencv_videostab.dylib -llibtbb.dylib
Cflags: -I${includedir_old} -I${includedir_new}@
I installed OpenCV 2.4.9 with Homebrew on OS X with TBB enabled. Using `pkg-config --libs --cflags opencv` results in an error because `libtbb.dylib` cannot be found. The problem is that the file exists (`/usr/local/lib/libtbb.dylib`) but the `Libs:` line should use `-ltbb` instead of `-llibtbb.dylib`. The first one works, the second fails with `ld: library not found for -llibtbb.dylib clang: error: linker command failed with exit code 1`.
```
## History
##### Vladislav Vinogradov on 2014-05-14 06:26
```
Alexander, could you check the issue?
- Assignee set to Alexander Smorkalov
- Status changed from New to Open
- Category set to build/install
```
##### Thorsten Ball on 2014-05-14 07:22
```
I'm happy to provide more information. If you need anything, just tell me.
```
| bug,auto-transferred,priority: normal,category: build/install,affected: 2.4 | low | Critical |
97,432,293 | opencv | Android setPictureSize is not respected in takePicture | Transferred from http://code.opencv.org/issues/3698
```
|| Renato Lorenzi on 2014-05-17 00:31
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: android
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: Any / Android
```
## Android setPictureSize is not respected in takePicture
```
I'm working with opencv4android and am using org.opencv.android.JavaCameraView class to display the image after processing. So far so good, but when I do a camera.takePicture () defined in the resolution parameters with parameters.setPictureSize (width, height) is not respected (get a lower resolution). Debugging the class code I found that the problem occurs because the call params.setRecordingHint (true) inside the initializeCamera method JavaCameraView class.
Looking JavaCameraView class review on github bug was discovered that caused the commit commit:98cb1dc
googled found a question where that commit had already caused problems [[http://answers.opencv.org/question/27600/camera-distorted-on-android-442/]]
the solution (bad) was to copy the code and remove the line
sorry my bad english
```
## History
##### Renato Lorenzi on 2014-05-17 16:08
```
- Status changed from New to Open
```
##### Alexander Smorkalov on 2014-09-27 14:28
```
What is your device? We approve pull request that turn this option off for Samsung Galaxy S2. It the patch applicable for you?
```
| bug,auto-transferred,priority: normal,affected: 2.4,platform: android | low | Critical |
97,432,334 | opencv | cv::Flann knnSearch distances | Transferred from http://code.opencv.org/issues/3702
```
|| Martin Beckett on 2014-05-20 21:06
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: flann
|| Tracker: Bug
|| Difficulty: Medium
|| PR:
|| Platform: x64 / Windows
```
## cv::Flann knnSearch distances
```
The returned distances from knnSearch don't represent the distance to the nearest neighbours returned in the index
It appears that they are the L1/L2 distance to the query points from the origin.
So with 2 query points and k=3 index | 1,2,3 || 2,3,4 | would return the same distances | d1,d1,d2 ||d2,d2,d2|
```
## History
##### Daniil Osokin on 2014-07-07 16:14
```
Hi Martin, thanks for response! Could you please submit a test to reproduce the bug, or, better, send us a pull request with the fix?
- Priority changed from High to Normal
- Assignee set to Martin Beckett
- Status changed from New to Incomplete
- Category set to flann
```
| bug,auto-transferred,priority: normal,affected: 2.4,category: flann | low | Critical |
97,432,367 | opencv | OpenCV 2.4.9 highgui multiple definitions of functions | Transferred from http://code.opencv.org/issues/3704
```
|| Michele Adduci on 2014-05-22 13:11
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: highgui-gui
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Windows
```
## OpenCV 2.4.9 highgui multiple definitions of functions
```
Hi
i'm trying to compile OpenCV 2.4.9 (downloaded from website) on a Windows 8.1 machine (x64) using CUDA 6.0 and OPENGL as enabled flags, under Visual Studio 2012. Everything goes fine except the generation of highgui.dll file, because i get the following error:
error LNK2005: cvCreateTrackbar already defined in window.obj C:\Frameworks\OpenCV\build\modules\highgui\window_w32.obj opencv_highgui
error LNK2005: cvCreateTrackbar2 already defined in window.obj C:\Frameworks\OpenCV\build\modules\highgui\window_w32.obj opencv_highgui
error LNK2005: cvDestroyAllWindows already defined in window.obj C:\Frameworks\OpenCV\build\modules\highgui\window_w32.obj opencv_highgui
error LNK2005: cvDestroyWindow already defined in window.obj C:\Frameworks\OpenCV\build\modules\highgui\window_w32.obj opencv_highgui
Error 5 error LNK2005: cvGetTrackbarPos already defined in window.obj C:\Frameworks\OpenCV\build\modules\highgui\window_w32.obj opencv_highgui
Error 6 error LNK2005: cvGetWindowHandle already defined in window.obj C:\Frameworks\OpenCV\build\modules\highgui\window_w32.obj opencv_highgui
Error 7 error LNK2005: cvGetWindowName already defined in window.obj C:\Frameworks\OpenCV\build\modules\highgui\window_w32.obj opencv_highgui
Error 8 error LNK2005: cvInitSystem already defined in window.obj C:\Frameworks\OpenCV\build\modules\highgui\window_w32.obj opencv_highgui
Error 9 error LNK2005: cvMoveWindow already defined in window.obj C:\Frameworks\OpenCV\build\modules\highgui\window_w32.obj opencv_highgui
Error 10 error LNK2005: cvNamedWindow already defined in window.obj C:\Frameworks\OpenCV\build\modules\highgui\window_w32.obj opencv_highgui
Error 11 error LNK2005: cvResizeWindow already defined in window.obj C:\Frameworks\OpenCV\build\modules\highgui\window_w32.obj opencv_highgui
Error 12 error LNK2005: cvSetMouseCallback already defined in window.obj C:\Frameworks\OpenCV\build\modules\highgui\window_w32.obj opencv_highgui
Error 13 error LNK2005: cvSetTrackbarPos already defined in window.obj C:\Frameworks\OpenCV\build\modules\highgui\window_w32.obj opencv_highgui
Error 14 error LNK2005: cvShowImage already defined in window.obj C:\Frameworks\OpenCV\build\modules\highgui\window_w32.obj opencv_highgui
Error 15 error LNK2005: cvStartWindowThread already defined in window.obj C:\Frameworks\OpenCV\build\modules\highgui\window_w32.obj opencv_highgui
Error 16 error LNK2005: cvWaitKey already defined in window.obj C:\Frameworks\OpenCV\build\modules\highgui\window_w32.obj opencv_highgui
Error 17 error LNK1169: one or more multiply defined symbols found C:\Frameworks\OpenCV\build\bin\Release\opencv_highgui249.dll 1 1 opencv_highgui
looking in the code, windows.cpp and windows_w32.cpp have the same definitions of same functions. How can it be solved?
```
## History
##### Michele Adduci on 2014-05-24 12:07
```
- Assignee set to Anatoly Baksheev
```
##### Dmitry Retinskiy on 2014-05-29 05:14
```
Roman, could you check this problem?
Thanks.
- Target version deleted (2.4.10)
- Priority changed from High to Normal
- Assignee changed from Anatoly Baksheev to Roman Donchenko
- Status changed from New to Open
```
| bug,auto-transferred,priority: normal,category: highgui-gui,affected: 2.4 | low | Critical |
97,432,570 | opencv | Algorithm::create doesn't work for BFMatcher | Transferred from http://code.opencv.org/issues/3718
```
|| Alexander Cheparukhin on 2014-05-27 15:29
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: core
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Linux
```
## Algorithm::create doesn't work for BFMatcher
```
cv::Algorithm::create<cv::DescriptorMatcher>("BFMatcher") returns a null pointer.
```
## History
##### Alexander Cheparukhin on 2014-05-27 15:39
```
This issue also occurs with for FlannBasedMatcher.
```
##### Dmitry Retinskiy on 2014-05-28 04:43
```
Hi Alexander,
here you can find possible explanation for this behavior: http://docs.opencv.org/modules/core/doc/basic_structures.html?highlight=algorithm%3A%3Acreate#algorithm-create. (See Note section there).
Please check if it helps and reply if the ticket can be closed.
Thanks.
- Assignee set to Alexander Cheparukhin
- Category set to core
```
| bug,auto-transferred,priority: normal,category: core,affected: 2.4 | low | Critical |
97,432,596 | opencv | Descriptor matchers don't serialize to FileStorage | Transferred from http://code.opencv.org/issues/3719
```
|| Alexander Cheparukhin on 2014-05-27 15:33
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: features2d
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Linux
```
## Descriptor matchers don't serialize to FileStorage
```
Both BFMatcher and FlannBasedMatcher do not serialize to FileStorage correctly, although they have read and write functions inherited from Algorithm class.
```
## History
##### Dmitry Retinskiy on 2014-05-28 04:48
```
Hi Alexander,
could you add more details to the description (a code snippet demonstrating the problem)?
Thanks.
- Assignee set to Alexander Cheparukhin
- Category set to features2d
```
| bug,auto-transferred,priority: normal,affected: 2.4,category: features2d | low | Critical |
97,432,618 | opencv | Flann LshIndexParams save and load does not work - SEGFAULT | Transferred from http://code.opencv.org/issues/3720
```
|| Puneet Jain on 2014-05-28 02:07
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: flann
|| Tracker: Bug
|| Difficulty: Easy
|| PR:
|| Platform: x64 / Linux
```
## Flann LshIndexParams save and load does not work - SEGFAULT
```
Test case: pass any binary descriptors (use AKAZE, ORB, or KAZE). This will through an error. I think /modules/flann/include/opencv2/flann/lsh_index.h
needs some fixes in loadIndex (load_value and save_value)
<pre>
void flannSaveAndLoadTest(Mat descriptor) {
// construct an index
flann::LshIndexParams lshIndexParams(20,10,2);
flann::Index kdtree(descriptor, lshIndexParams, cvflann::FLANN_DIST_HAMMING);
// save descriptors to file
writeMatToFile(descriptor, "before_index_save.dat");
// save index to file
kdtree.save("index.dat");
// load index from file
Mat mat = Mat(descriptor.rows, descriptor.cols, CV_8UC1);
flann::Index kdtree1(mat, flann::SavedIndexParams("index.dat"), cvflann::FLANN_DIST_HAMMING);
// save loaded mat
writeMatToFile(mat, "after_index_save.dat");
// search in this tree
Mat dists = Mat(descriptor.rows, 2, CV_32S);
Mat indices = Mat(descriptor.rows, 3, CV_32S);
kdtree1.knnSearch(descriptor, indices, dists, 2);
}
</pre>
```
## History
##### Dmitry Retinskiy on 2014-05-28 04:25
```
Hi Puneet Jain,
it would be appreciated if you could suggest a fix for the reported issue.
(see here for instructions on how to do it: http://www.code.opencv.org/projects/opencv/wiki/How_to_contribute)
- Assignee set to Puneet Jain
- Status changed from New to Open
- Category set to flann
```
##### Puneet Jain on 2014-05-28 23:54
```
Hi Dmitry Retinskiy,
After spending several hours, I was able to triage the issue. The issue is not in save/load of indices in LSH but in modules/flann/include/opencv2/flann/lsh_tables.h. The problem is as follows:
lsh_index.h when loads saved indices data, it uses add() method defined in lsh_tables.
<pre>
/** Add a feature to the table
* @param value the value to store for that feature
* @param feature the feature itself
*/
void add(unsigned int value, const ElementType* feature)
{
// Add the value to the corresponding bucket
BucketKey key = (lsh::BucketKey)getKey(feature);
switch (speed_level_) {
case kArray:
// That means we get the buckets from an array
buckets_speed_[key].push_back(value);
break;
case kBitsetHash:
// That means we can check the bitset for the presence of a key
key_bitset_.set(key);
buckets_space_[key].push_back(value);
break;
case kHash:
{
// That means we have to check for the hash table for the presence of a key
buckets_space_[key].push_back(value);
break;
}
}
}
</pre>
There is something called speed_level which is decided based on the your amount of data/ or some other parameter. The problem occurs when addition is done at one speed and retrieval (getBucketFromKey) at a different speed. In my case, the addition was happening at kHash speed, but retrieval at kBitHash speed.
The quick fix for this is to make kHash as default speed for everthing:
<pre>
inline const Bucket* getBucketFromKey(BucketKey key) const
{
// Generate other buckets
switch (speed_level_) {
case kArray:
/* // That means we get the buckets from an array
return &buckets_speed_[key];
*/
{
// That means we have to check for the hash table for the presence of a key
BucketsSpace::const_iterator bucket_it, bucket_end = buckets_space_.end();
bucket_it = buckets_space_.find(key);
// Stop here if that bucket does not exist
if (bucket_it == bucket_end) return 0;
else return &bucket_it->second;
}
break;
case kBitsetHash:
/* // That means we can check the bitset for the presence of a key
if (key_bitset_.test(key)) return &buckets_space_.find(key)->second;
else return 0;
*/
{
// That means we have to check for the hash table for the presence of a key
BucketsSpace::const_iterator bucket_it, bucket_end = buckets_space_.end();
bucket_it = buckets_space_.find(key);
// Stop here if that bucket does not exist
if (bucket_it == bucket_end) return 0;
else return &bucket_it->second;
}
break;
case kHash:
{
// That means we have to check for the hash table for the presence of a key
BucketsSpace::const_iterator bucket_it, bucket_end = buckets_space_.end();
bucket_it = buckets_space_.find(key);
// Stop here if that bucket does not exist
if (bucket_it == bucket_end) return 0;
else return &bucket_it->second;
break;
}
}
return 0;
}
</pre>
The commented part of code is the current code in OpenCV. I have pasted kHash in all three cases. It seem to work fine now. (I am not certain about correctness though). Hope someone can pick it from here and fix this issue.
Thanks
Puneet
```
| bug,auto-transferred,priority: normal,affected: 3.4,category: flann | low | Critical |
97,432,650 | opencv | Need Flann Index to support vector<Mat> and return imgIdx | Transferred from http://code.opencv.org/issues/3721
```
|| Puneet Jain on 2014-05-28 08:10
|| Priority: Normal
|| Affected: None
|| Category: flann
|| Tracker: Feature
|| Difficulty:
|| PR:
|| Platform: None / None
```
## Need Flann Index to support vector<Mat> and return imgIdx
```
Hello,
I am trying to construct flann based indexes for multiple images. Ideally, I would like to supply a extracted descriptors of all images in the form of vector<Mat> to Index :
vector<Mat> descriptors;
flann::Index kdtree(descriptors, indexParams);
Once the index is built, I can dump in into disk like:
kdtree.save(outputFile);
This index, I should be able to load back and use it for the matching. I would also like to request knnSearch to return original image index:
kdtree.knnSearch(descriptors, indices, dists, 2);
where indices should be 3 values: <imgIdx, match1, match2>
All of this is supported in Lowe's flann library but somehow getting lost in OpenCV.
```
## History
##### Dmitry Retinskiy on 2014-05-29 07:54
```
Hi Puneet,
do you think you could add such a functionality to opencv?
```
##### Puneet Jain on 2014-05-30 17:19
```
Hi Dmitry Retinskiy,
I thought about it but it seem to be a bit complex and I am sorry I don't have enough cycles to explore this further. For now, I am maintaining by own range file such that I can identify imageIdx from the matched index in Mat.
This is what I am doing:
superMat = hconcat(vector<Mat>);
// dump indices range to a file Mat1: 0-a, Mat2: a+1-b, .....
flann::Index kdtree(superMat, indexParams);
kdtree.knnSearch(descriptors, indices, dists, 2);
==> this gives me indices in superMat
==> identify image using dumped indices
```
##### Alexander Karsakov on 2014-06-10 18:45
```
- Status changed from New to Open
- Category set to flann
```
| auto-transferred,priority: normal,feature,category: flann | low | Minor |
97,432,878 | opencv | CUDA exceptions in OpenCV GPU module in multi-threaded version | Transferred from http://code.opencv.org/issues/3738
```
|| T Abdullah on 2014-06-04 16:05
|| Priority: Normal
|| Affected: 2.4.0 - 2.4.8
|| Category: gpu (cuda)
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Windows
```
## CUDA exceptions in OpenCV GPU module in multi-threaded version
```
Hi,
Dear all,
I am trying to run a multi-threaded application using OpenCV GPU module. The application is developed in Visual Studio 2012 professionl + CUDA 5.5 + OpenCV 2.4.6 and uses pthreads for multi-threading. I built OpenCV 2.4.6 for this environment.
My application executes with two threads and there are no exceptions. I start getting a number of CUDA related exceptions with 3 or more that three threads. Single application of multiple application instances. It does not matter
am observing the same behavior.
OpenCV Error: Gpu API call (NCV Assertion Failed: cudaError_t=4, file=C:/opencv-2.4.6/modules/gpu/src/nvidia/NCVHaarObjectDetection.cu, line=1306) in unknown function, file C:\opencv-2.4.6\modules\gpu\src\cascadeclassifier.cpp, line 173
OpenCV Error: Gpu API call (unspecified launch failure) in unknown function, file C:\opencv-2.4.6\modules\core\src\gpumat.cpp, line 1196
OpenCV Error: Gpu API call (NCV Assertion Failed: cudaError_t=4, file=C:/opencv-2.4.6/modules/gpu/src/nvidia/NCVHaarObjectDetection.cu, line=1306) in unknown function, file C:\opencv-2.4.6\modules\gpu\src\cascadeclassifier.cpp, line 173
OpenCV Error: Gpu API call (NCV Assertion Failed: cudaError_t=4, file=C:/opencv-2.4.6/modules/gpu/src/nvidia/NCVHaarObjectDetection.cu, line=1523) in unknown function, file C:\opencv-2.4.6\modules\gpu\src\cascadeclassifier.cpp, line 173
OpenCV Error: Gpu API call (unspecified launch failure) in unknown function, file C:\opencv-2.4.6\modules\core\src\gpumat.cpp, line 1415
OpenCV Error: Gpu API call (NCV Assertion Failed: cudaError_t=4, file=C:/opencv-2.4.6/modules/gpu/src/nvidia/NCVHaarObjectDetection.cu, line=1523) in unknown function, file C:\opencv-2.4.6\modules\gpu\src\cascadeclassifier.cpp, line 173
OpenCV Error: Gpu API call (NCV Assertion Failed: retcode=2, file=C:/opencv-2.4.6/modules/gpu/src/nvidia/NCVHaarObjectDetection.cu, line=1768) in unknown function, file C:\opencv-2.4.6\modules\gpu\src\cascadeclassifier.cpp, line 173
OpenCV Error: Gpu API call (NCV Assertion Failed: retcode=5, file=C:/opencv-2.4.6/modules/gpu/src/nvidia/NCVHaarObjectDetection.cu, line=1575) in unknown function, file C:\opencv-2.4.6\modules\gpu\src\cascadeclassifier.cpp, line 173
I can provide more details if needed.
I would also welcome any alternate suggestions for solving these exceptions as requested by me in this thread. [http://answers.opencv.org/question/34578/best-practices-for-opencv-gpu-module-in-multi/]
```
## History
##### Alexander Karsakov on 2014-06-09 14:04
```
Vlad, could you please take a look?
- Assignee set to Vladislav Vinogradov
- Status changed from New to Open
- Category set to gpu (cuda)
```
##### T Abdullah on 2014-07-11 16:46
```
Hi Vladislav/alexander,
It has been more than a month after submitting the issue. I am still facing the crash. I can provide more details if needed to solve my problem.
Regards,
Tariq Abdullah
```
| bug,auto-transferred,priority: normal,affected: 2.4,category: gpu/cuda (contrib) | low | Critical |
97,432,902 | opencv | Add support for true in-place convolution to filter2D | Transferred from http://code.opencv.org/issues/3741
```
|| Goncalo Lopes on 2014-06-05 10:15
|| Priority: Normal
|| Affected: None
|| Category: imgproc, video
|| Tracker: Feature
|| Difficulty:
|| PR:
|| Platform: None / None
```
## Add support for true in-place convolution to filter2D
```
The documentation for filter2D claims that "in-place operation is supported.". However, this is not really the case, and OpenCV seems to create a temporary copy somewhere to serve as input to the filter. This is unfortunate, since it precludes the possibility of designing true IIR filters using this operator (where you want the previous output to affect the next input).
It would be nice to add a flag to allow specifying true in-place convolution.
```
## History
##### Daniil Osokin on 2014-07-07 16:42
```
- Category set to imgproc, video
```
| auto-transferred,priority: normal,feature,category: imgproc,category: video | low | Minor |
97,432,972 | opencv | gpu:HOGDescriptor cannot support two hog descriptor with different parameters | Transferred from http://code.opencv.org/issues/3748
```
|| Zhifei Fang on 2014-06-11 19:45
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: gpu (cuda)
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Windows
```
## gpu:HOGDescriptor cannot support two hog descriptor with different parameters
```
In namespace hog, there are some constants. Defined in hog.cu.
hog.cu line 60~68 :
__constant__ int cnbins;
__constant__ int cblock_stride_x;
__constant__ int cblock_stride_y;
__constant__ int cnblocks_win_x;
__constant__ int cnblocks_win_y;
__constant__ int cblock_hist_size;
__constant__ int cblock_hist_size_2up;
__constant__ int cdescr_size;
__constant__ int cdescr_width;
These constant assigned by a function called *set_up_constants*
hog.cu line 90:
void set_up_constants(int nbins, int block_stride_x, int block_stride_y,
int nblocks_win_x, int nblocks_win_y)
In the hog.cpp (gpu module)
The constructor of gpu:HOGDescriptor (line 132) will call *set_up_constants*
so if user want to set two HOGDescriptor with different parameters, that may cause a problem.
The second will rewrite these global, that make the first HOGDescriptor cannot work in the proper way.
I will attach a test file later.
```
## History
##### Daniil Osokin on 2014-07-07 15:06
```
Hi Zhifei, will you provide any tests?
- Priority changed from High to Normal
- Assignee set to Zhifei Fang
- Status changed from New to Incomplete
- Category set to gpu (cuda)
```
##### Tran Duong on 2014-10-25 09:38
```
I have same problem. I create two model files for detect motobike and car with different parameters. If I use only motobike model to detect, that's no problem (FPS higher than 30). But when I use two model to detect, the program detect very slowly (FPS less 1).
Note: this bug don't occur on CPU mode
```
##### Huy Le on 2014-10-27 05:34
```
I got the same problem. Please fix it in version 3.0.
```
| bug,auto-transferred,priority: normal,affected: 2.4,category: gpu/cuda (contrib) | low | Critical |
97,433,096 | opencv | SparseMatIterator_<> build error | Transferred from http://code.opencv.org/issues/3761
```
|| Marc Rollins on 2014-06-19 21:10
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: core
|| Tracker: Bug
|| Difficulty: Easy
|| PR:
|| Platform: Any / Any
```
## SparseMatIterator_<> build error
```
Compilation fails with errors:
<pre>
/usr/local/include/opencv2/core/mat.hpp:2611:23: No viable conversion from 'cv::SparseMatIterator_<float>' to 'cv::SparseMatIterator'
/usr/local/include/opencv2/core/mat.hpp:2569:12: error: no viable conversion from 'cv::SparseMatConstIterator' to 'cv::SparseMatConstIterator_<float>'
</pre>
when iterating a @SparseMat_@ using templated @SparseMatIterator_@ in a @for@ loop and using post-increment @operator++@.
Compiler invocation:
<pre>
clang++ sparseMat-repro.cpp -I/usr/local/include -L/usr/local/lib
</pre>
A fully reproducible sample is attached.
This error occurs compiling with both g++ and clang++.
```
## History
| bug,auto-transferred,priority: normal,category: core,affected: 2.4 | low | Critical |
97,433,131 | opencv | opencv2/flann/lsh_table.h 'unordered_map' file not found | Transferred from http://code.opencv.org/issues/3773
```
|| Liu Xiao on 2014-06-27 13:16
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: flann
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Mac OSX
```
## opencv2/flann/lsh_table.h 'unordered_map' file not found
```
When I try to use opencv in my project, it has error below:
opencv2/flann/lsh_table.h 'unordered_map' file not found
```
## History
##### Ilya Lavrenov on 2014-07-04 08:14
```
Hi Hoa, could you include http://www.cplusplus.com/reference/unordered_map/unordered_map/ in the appropriate header file, test the fix and make pull-request to https://github.com/Itseez/opencv?
```
| auto-transferred,category: build/install,affected: 2.4,category: flann,incomplete | low | Critical |
97,433,192 | opencv | Adding BinBoost to OpenCV without a "create" implementation | Transferred from http://code.opencv.org/issues/3776
```
|| Gil Levi on 2014-06-27 16:25
|| Priority: Normal
|| Affected: None
|| Category: features2d
|| Tracker: Feature
|| Difficulty:
|| PR:
|| Platform: None / None
```
## Adding BinBoost to OpenCV without a "create" implementation
```
Hi,
I'm trying to add the BinBoost descriptor to OpenCV. The sources can be found here:
http://cvlab.epfl.ch/research/detect/binboost
It's really straightforward, as the authors already implemented the DescriptorExtractor class.
The problem is that the constructors are dependent of certain binary files as input. They use them to initialize their inner structures. So one can easily construct a BinBoostDescriptorExtractor as
> BinBoostDescriptorExtractor BinBoostDescriptorExtractorInstance("D:\\OpenCV_2_4_9\\opencv\\sources\\modules\\features2d\\src\\binboost_256.bin");
But one cannot use the simpler "create" command as:
> Ptr<DescriptorExtractor> descriptor = DescriptorExtractor::create("BinBoost");
I already asked in the OpenCV Q&A forum, and no one could provide a solution:
http://answers.opencv.org/question/35272/problems-in-adding-a-new-descriptor-to-opencv/
Moderators - will you be willing to a accept a family of 3 new descriptors with the downside that they don't have a "algorithm::create" implementation? I must note that in my experiments, BinBoost is state-of-the-art among the binary descriptors.
Thanks,
Gil
```
## History
| auto-transferred,priority: normal,feature,category: features2d | low | Minor |
97,433,222 | opencv | Can't pass precalculated pyramids to calcOpticalFlowPyrLK in Python | Transferred from http://code.opencv.org/issues/3784
```
|| Magnus Linderoth on 2014-06-30 17:26
|| Priority: Normal
|| Affected: 2.4.0 - 2.4.8
|| Category: python bindings
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Mac OSX
```
## Can't pass precalculated pyramids to calcOpticalFlowPyrLK in Python
```
In Python I get an error when I try to use the output from buildOpticalFlowPyramid as input to calcOpticalFlowPyrLK.
An example is given here:
>>> im1 = np.uint8(np.random.rand(64,96)*255)
>>> im2 = np.uint8(np.random.rand(64,96)*255)
>>> pyr1 = cv2.buildOpticalFlowPyramid(im1, (7,7), 2)
>>> pyr2 = cv2.buildOpticalFlowPyramid(im2, (7,7), 2)
>>> cv2.calcOpticalFlowPyrLK(pyr1[1], pyr2[1], np.float32((42,42)))
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
TypeError: prevImg is not a numpy array, neither a scalar
The problem was previously described here:
http://answers.opencv.org/question/315/cant-use-pyramids-in-calcopticalflowpyrlk-python/
The problem seems to be related to an issue reported for the Java interface:
http://code.opencv.org/issues/3508
```
## History
| auto-transferred,priority: normal,category: python bindings,category: video,affected: 2.4,RFC | low | Critical |
97,433,314 | opencv | install opencv-2.4-9 on centos 6.5 with cuda 5.5 support | Transferred from http://code.opencv.org/issues/3794
```
|| ricardo roman on 2014-07-04 15:48
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: build/install
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Linux
```
## install opencv-2.4-9 on centos 6.5 with cuda 5.5 support
```
Hello. I have tried to get support on this from forums and wikis to no avail. I guess then its a bug.
I can't seem to install opencv 2.4.9 on centos 6.5 with cuda 5.5 support. this is my cmake output:
[root@cadejos-0 opencv]# cmake -D CMAKE_BUILD_TYPE=RELEASE -D CMAKE_INSTALL_PREFIX=/opt/opencv -D WITH_CUDA=ON /opt/opencv-2.4.9/opencv/
-- Detected version of GNU GCC: 44 (404)
-- Looking for linux/videodev.h
-- Looking for linux/videodev.h - found
-- Looking for linux/videodev2.h
-- Looking for linux/videodev2.h - found
-- Looking for sys/videoio.h
-- Looking for sys/videoio.h - not found
-- Looking for libavformat/avformat.h
-- Looking for libavformat/avformat.h - found
-- Looking for ffmpeg/avformat.h
-- Looking for ffmpeg/avformat.h - not found
Traceback (most recent call last):
File "<string>", line 1, in <module>
ImportError: No module named numpy.distutils
-- Found apache ant 1.7.1: /usr/bin/ant
--
-- General configuration for OpenCV 2.4.8.2 =====================================
-- Version control: 2.4.8.2-dirty
--
-- Platform:
-- Host: Linux 2.6.32-358.18.1.el6.x86_64 x86_64
-- CMake: 2.6.4
-- CMake generator: Unix Makefiles
-- CMake build tool: /usr/bin/gmake
-- Configuration: RELEASE
--
-- C/C++:
-- Built as dynamic libs?: YES
-- C++ Compiler: /usr/bin/c++ (ver 4.4.7)
-- C++ flags (Release): -fsigned-char -W -Wall -Werror=return-type -Werror=address -Werror=sequence-point -Wformat -Werror=format-security -Wmissing-declarations -Wundef -Winit-self -Wpointer-arith -Wshadow -Wsign-promo -fdiagnostics-show-option -Wno-long-long -pthread -fomit-frame-pointer -msse -msse2 -msse3 -ffunction-sections -O3 -DNDEBUG -DNDEBUG
-- C++ flags (Debug): -fsigned-char -W -Wall -Werror=return-type -Werror=address -Werror=sequence-point -Wformat -Werror=format-security -Wmissing-declarations -Wundef -Winit-self -Wpointer-arith -Wshadow -Wsign-promo -fdiagnostics-show-option -Wno-long-long -pthread -fomit-frame-pointer -msse -msse2 -msse3 -ffunction-sections -g -O0 -DDEBUG -D_DEBUG
-- C Compiler: /usr/bin/gcc
-- C flags (Release): -fsigned-char -W -Wall -Werror=return-type -Werror=address -Werror=sequence-point -Wformat -Werror=format-security -Wmissing-declarations -Wmissing-prototypes -Wstrict-prototypes -Wundef -Winit-self -Wpointer-arith -Wshadow -fdiagnostics-show-option -Wno-long-long -pthread -fomit-frame-pointer -msse -msse2 -msse3 -ffunction-sections -O3 -DNDEBUG -DNDEBUG
-- C flags (Debug): -fsigned-char -W -Wall -Werror=return-type -Werror=address -Werror=sequence-point -Wformat -Werror=format-security -Wmissing-declarations -Wmissing-prototypes -Wstrict-prototypes -Wundef -Winit-self -Wpointer-arith -Wshadow -fdiagnostics-show-option -Wno-long-long -pthread -fomit-frame-pointer -msse -msse2 -msse3 -ffunction-sections -g -O0 -DDEBUG -D_DEBUG
-- Linker flags (Release):
-- Linker flags (Debug):
-- Precompiled headers: YES
--
-- OpenCV modules:
-- To be built: core flann imgproc highgui features2d calib3d ml video legacy objdetect photo gpu ocl nonfree contrib stitching superres ts videostab
-- Disabled: world
-- Disabled by dependency: -
-- Unavailable: androidcamera dynamicuda java python
--
-- GUI:
-- QT: NO
-- GTK+ 2.x: YES (ver 2.20.1)
-- GThread : YES (ver 2.26.1)
-- GtkGlExt: NO
-- OpenGL support: NO
--
-- Media I/O:
-- ZLib: /usr/lib64/libz.so (ver 1.2.3)
-- JPEG: /usr/lib64/libjpeg.so (ver )
-- PNG: /usr/lib64/libpng.so (ver 1.2.49)
-- TIFF: /usr/lib64/libtiff.so (ver 42 - 3.9.4)
-- JPEG 2000: /usr/lib64/libjasper.so (ver 1.900.1)
-- OpenEXR: build (ver 1.7.1)
--
-- Video I/O:
-- DC1394 1.x: NO
-- DC1394 2.x: YES (ver 2.1.2)
-- FFMPEG: YES
-- codec: YES (ver 52.72.2)
-- format: YES (ver 52.64.2)
-- util: YES (ver 50.15.1)
-- swscale: YES (ver 0.11.0)
-- gentoo-style: YES
-- GStreamer:
-- base: YES (ver 0.10.29)
-- app: YES (ver 0.10.29)
-- video: YES (ver 0.10.29)
-- OpenNI: NO
-- OpenNI PrimeSensor Modules: NO
-- PvAPI: NO
-- GigEVisionSDK: NO
-- UniCap: NO
-- UniCap ucil: NO
-- V4L/V4L2: Using libv4l (ver 0.6.3)
-- XIMEA: NO
-- Xine: NO
--
-- Other third-party libraries:
-- Use IPP: NO
-- Use Eigen: NO
-- Use TBB: NO
-- Use OpenMP: NO
-- Use GCD NO
-- Use Concurrency NO
-- Use C=: NO
-- Use Cuda: NO
-- Use OpenCL: YES
--
-- OpenCL:
-- Version: dynamic
-- Include path: /opt/opencv-2.4.9/opencv/3rdparty/include/opencl/1.2
-- Use AMD FFT: NO
-- Use AMD BLAS: NO
--
-- Python:
-- Interpreter: /usr/bin/python2.6 (ver 2.6.6)
--
-- Java:
-- ant: /usr/bin/ant (ver 1.7.1)
-- JNI: /usr/lib/jvm/java/include /usr/lib/jvm/java/include/linux /usr/lib/jvm/java/include
-- Java tests: NO
--
-- Documentation:
-- Build Documentation: NO
-- Sphinx: NO
-- PdfLaTeX compiler: /usr/bin/pdflatex
--
-- Tests and samples:
-- Tests: YES
-- Performance tests: YES
-- C/C++ Examples: NO
--
-- Install path: /opt/opencv
--
-- cvconfig.h is in: /opt/opencv-2.4.9/opencv
-- -----------------------------------------------------------------
--
WARNINGThe source directory is the same as binary directory. "make clean" may damage the source tree
-- Configuring done
-- Generating done
-- Build files have been written to: /opt/opencv-2.4.9/opencv
I have cuda working:
[root@cadejos-0 opencv]# /usr/local/cuda-5.5/NVIDIA_CUDA-5.5_Samples/1_Utilities/deviceQuery/deviceQuery
/usr/local/cuda-5.5/NVIDIA_CUDA-5.5_Samples/1_Utilities/deviceQuery/deviceQuery Starting...
CUDA Device Query (Runtime API) version (CUDART static linking)
Detected 2 CUDA Capable device(s)
Device 0: "Tesla M1060"
CUDA Driver Version / Runtime Version 5.5 / 5.5
CUDA Capability Major/Minor version number: 1.3
Total amount of global memory: 4096 MBytes (4294770688 bytes)
(30) Multiprocessors, ( 8) CUDA Cores/MP: 240 CUDA Cores
GPU Clock rate: 1296 MHz (1.30 GHz)
Memory Clock rate: 800 Mhz
Memory Bus Width: 512-bit
Maximum Texture Dimension Size (x,y,z) 1D=(8192), 2D=(65536, 32768), 3D=(2048, 2048, 2048)
Maximum Layered 1D Texture Size, (num) layers 1D=(8192), 512 layers
Maximum Layered 2D Texture Size, (num) layers 2D=(8192, 8192), 512 layers
Total amount of constant memory: 65536 bytes
Total amount of shared memory per block: 16384 bytes
Total number of registers available per block: 16384
Warp size: 32
Maximum number of threads per multiprocessor: 1024
Maximum number of threads per block: 512
Max dimension size of a thread block (x,y,z): (512, 512, 64)
Max dimension size of a grid size (x,y,z): (65535, 65535, 1)
Maximum memory pitch: 2147483647 bytes
Texture alignment: 256 bytes
Concurrent copy and kernel execution: Yes with 1 copy engine(s)
Run time limit on kernels: No
Integrated GPU sharing Host Memory: No
Support host page-locked memory mapping: Yes
Alignment requirement for Surfaces: Yes
Device has ECC support: Disabled
Device supports Unified Addressing (UVA): No
Device PCI Bus ID / PCI location ID: 2 / 0
Compute Mode:
< Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >
Device 1: "Tesla M1060"
CUDA Driver Version / Runtime Version 5.5 / 5.5
CUDA Capability Major/Minor version number: 1.3
Total amount of global memory: 4096 MBytes (4294770688 bytes)
(30) Multiprocessors, ( 8) CUDA Cores/MP: 240 CUDA Cores
GPU Clock rate: 1296 MHz (1.30 GHz)
Memory Clock rate: 800 Mhz
Memory Bus Width: 512-bit
Maximum Texture Dimension Size (x,y,z) 1D=(8192), 2D=(65536, 32768), 3D=(2048, 2048, 2048)
Maximum Layered 1D Texture Size, (num) layers 1D=(8192), 512 layers
Maximum Layered 2D Texture Size, (num) layers 2D=(8192, 8192), 512 layers
Total amount of constant memory: 65536 bytes
Total amount of shared memory per block: 16384 bytes
Total number of registers available per block: 16384
Warp size: 32
Maximum number of threads per multiprocessor: 1024
Maximum number of threads per block: 512
Max dimension size of a thread block (x,y,z): (512, 512, 64)
Max dimension size of a grid size (x,y,z): (65535, 65535, 1)
Maximum memory pitch: 2147483647 bytes
Texture alignment: 256 bytes
Concurrent copy and kernel execution: Yes with 1 copy engine(s)
Run time limit on kernels: No
Integrated GPU sharing Host Memory: No
Support host page-locked memory mapping: Yes
Alignment requirement for Surfaces: Yes
Device has ECC support: Disabled
Device supports Unified Addressing (UVA): No
Device PCI Bus ID / PCI location ID: 3 / 0
Compute Mode:
< Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >
deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 5.5, CUDA Runtime Version = 5.5, NumDevs = 2, Device0 = Tesla M1060, Device1 = Tesla M1060
Result = PASS
[root@cadejos-0 opencv]#
But somehow the cmake doesnt find them. These are my variables:
[root@cadejos-0 opencv]# export
declare -x CVS_RSH="ssh"
declare -x G_BROKEN_FILENAMES="1"
declare -x HISTCONTROL="ignoredups"
declare -x HISTSIZE="1000"
declare -x HOME="/root"
declare -x HOSTNAME="cadejos-0"
declare -x LANG="en_US.UTF-8"
declare -x LD_LIBRARY_PATH="/usr/local/cuda-5.5/lib64:/usr/local/cuda-5.5/lib:/lib64:/lib:/usr/lib:/usr/lib64:"
declare -x LESSOPEN="|/usr/bin/lesspipe.sh %s"
declare -x LOADEDMODULES=""
declare -x LOGNAME="root"
declare -x LS_COLORS="rs=0:di=01;34:ln=01;36:mh=00:pi=40;33:so=01;35:do=01;35:bd=40;33;01:cd=40;33;01:or=40;31;01:mi=01;05;37;41:su=37;41:sg=30;43:ca=30;41:tw=30;42:ow=34;42:st=37;44:ex=01;32:*.tar=01;31:*.tgz=01;31:*.arj=01;31:*.taz=01;31:*.lzh=01;31:*.lzma=01;31:*.tlz=01;31:*.txz=01;31:*.zip=01;31:*.z=01;31:*.Z=01;31:*.dz=01;31:*.gz=01;31:*.lz=01;31:*.xz=01;31:*.bz2=01;31:*.tbz=01;31:*.tbz2=01;31:*.bz=01;31:*.tz=01;31:*.deb=01;31:*.rpm=01;31:*.jar=01;31:*.rar=01;31:*.ace=01;31:*.zoo=01;31:*.cpio=01;31:*.7z=01;31:*.rz=01;31:*.jpg=01;35:*.jpeg=01;35:*.gif=01;35:*.bmp=01;35:*.pbm=01;35:*.pgm=01;35:*.ppm=01;35:*.tga=01;35:*.xbm=01;35:*.xpm=01;35:*.tif=01;35:*.tiff=01;35:*.png=01;35:*.svg=01;35:*.svgz=01;35:*.mng=01;35:*.pcx=01;35:*.mov=01;35:*.mpg=01;35:*.mpeg=01;35:*.m2v=01;35:*.mkv=01;35:*.ogm=01;35:*.mp4=01;35:*.m4v=01;35:*.mp4v=01;35:*.vob=01;35:*.qt=01;35:*.nuv=01;35:*.wmv=01;35:*.asf=01;35:*.rm=01;35:*.rmvb=01;35:*.flc=01;35:*.avi=01;35:*.fli=01;35:*.flv=01;35:*.gl=01;35:*.dl=01;35:*.xcf=01;35:*.xwd=01;35:*.yuv=01;35:*.cgm=01;35:*.emf=01;35:*.axv=01;35:*.anx=01;35:*.ogv=01;35:*.ogx=01;35:*.aac=01;36:*.au=01;36:*.flac=01;36:*.mid=01;36:*.midi=01;36:*.mka=01;36:*.mp3=01;36:*.mpc=01;36:*.ogg=01;36:*.ra=01;36:*.wav=01;36:*.axa=01;36:*.oga=01;36:*.spx=01;36:*.xspf=01;36:"
declare -x MAIL="/var/spool/mail/root"
declare -x MODULEPATH="/usr/share/Modules/modulefiles:/etc/modulefiles"
declare -x MODULESHOME="/usr/share/Modules"
declare -x OLDPWD="/opt"
declare -x PATH="/usr/local/cuda-5.5/bin:/opt/mothur/:/opt/usearch/:/usr/local/sbin:/usr/local/bin:/sbin:/bin:/usr/sbin:/usr/bin:/root/bin"
declare -x PWD="/opt/opencv"
declare -x SHELL="/bin/bash"
declare -x SHLVL="1"
declare -x SSH_CLIENT="10.0.0.1 48436 22"
declare -x SSH_CONNECTION="10.0.0.1 48436 10.0.0.10 22"
declare -x SSH_TTY="/dev/pts/0"
declare -x TERM="xterm"
declare -x USER="root"
What am I doing wrong, and how can i solve it?
Thank you
```
## History
| bug,auto-transferred,priority: normal,category: build/install,affected: 2.4 | low | Critical |
97,433,393 | opencv | Fixing user guide tutorials | Transferred from http://code.opencv.org/issues/3798
```
|| Steven Puttemans on 2014-07-07 13:08
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: documentation
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: Any / Any
```
## Fixing user guide tutorials
```
http://docs.opencv.org/doc/user_guide/ug_features2d.html
Looking at user guides like this, it seems to me they should better be fixed or removed all together. The guide does nothing extra compared to the actual sample provided... the minimal comments should just be added in comments inside the sample I guess.
```
## History
| bug,auto-transferred,priority: normal,category: documentation,affected: 2.4 | low | Critical |
97,433,428 | opencv | Qt display keepratio mode | Transferred from http://code.opencv.org/issues/3800
```
|| Florent Perrocheau on 2014-07-08 11:39
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: highgui-gui
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x86 / Windows
```
## Qt display keepratio mode
```
Hy,
since I work with 2.4.9, the keep ratio display mode with Qt doesn't work well.
The output image size is mostly ok but the image data is not rescaled.
Some time the output image is not resized.
!http://s28.postimg.org/pr0ufw0lp/bug_opencv_qt_display.png!
OpenCV 2.4.9
Qt 5.3.1
mingw32 4.8.2
```
## History
##### Florent Perrocheau on 2014-07-08 13:28
```
Same displays on freeratio mode.
!http://s28.postimg.org/465gqoc25/bug_opencv_qt_display_2.png!
```
##### Florent Perrocheau on 2014-07-08 17:58
```
After some tests the problem doesn't come with opencv 2.4.9 but with the upgrade to Qt 5.
2.4.9 and 2.4.7 versions compile with Qt 4.8.5 and mingw32 4.6.2 work great.
2.4.9, 2.4.8 and 2.4.7 versions compile with Qt 5.3.1 and mingw32 4.8.2 are buggy.
I will try to isolate compiler and Qt version as far it possible (don't success to mix Qt5 and mingw 4.6.2).
Anyone have already encounter this problem?
Edit :
The bug only appears when OpenCV is used on a Qt application with the last version of Qt (5.3.1) and wathever the version of Qt used to compile OpenCV (testing 5.2.0 and 5.3.1) and the compiler used (testing mingw 4.8.2, 4.8.0 and 4.7).
```
##### Dmitry Retinskiy on 2014-09-17 09:50
```
Florent,
according to your investigation it looks like the problem is in Qt version, not in OpenCV.
If you agree, I'd suggest to close the ticket.
```
| bug,auto-transferred,priority: normal,category: highgui-gui,affected: 2.4 | low | Critical |
97,433,460 | opencv | VideoCapture behaves different for jpeg and png sequences | Transferred from http://code.opencv.org/issues/3806
```
|| Philipp Hasper on 2014-07-11 15:04
|| Priority: Normal
|| Affected: 2.4.0 - 2.4.8
|| Category: highgui-camera
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: Any / Any
```
## VideoCapture behaves different for jpeg and png sequences
```
the VideoCapture class is able to read image sequences specified something like
# _%05d.jpg_
# _%05d.png_
# _00000.png_
but not _00000.jpg_.
This is because in the first three cases
@result = cvCreateFileCapture_FFMPEG_proxy (filename);@ (cap.cpp, line 372)
fails and another handler jumps in successfully
But in the fourth case
@result = cvCreateFileCapture_FFMPEG_proxy (filename);@
is successful(cap.cpp, line 372).
And the FFMPEG proxy is not able to process file sequences specified by their first file (at least this _seems_ to be the problem).
How can we prevent the proxy to accept the sequence although not being able to process it? The most simple solution would be to move the FFMPEG proxy to the bottom but I don't know if this would break anything else.
```
## History
##### Dmitry Retinskiy on 2014-09-17 09:46
```
Philipp, you could try your suggestion using pull request (http://www.code.opencv.org/projects/opencv/wiki/How_to_contribute).
It will be first automatically tested (http://pullrequest.opencv.org/) and then reviewed.
- Assignee set to Philipp Hasper
```
##### Dmitry Retinskiy on 2014-09-17 09:46
```
- Status changed from New to Open
```
| bug,auto-transferred,priority: low,affected: 2.4,category: videoio(camera) | low | Critical |
97,433,602 | opencv | New class to be added to OpenCV | Transferred from http://code.opencv.org/issues/3810
```
|| Robby Longhorn on 2014-07-13 18:34
|| Priority: Normal
|| Affected: None
|| Category: highgui-images
|| Tracker: Feature
|| Difficulty:
|| PR:
|| Platform: None / None
```
## New class to be added to OpenCV
```
I wrote a class that is very usefull for all Windows users of OpenCV.
I ask herewith to add this class to the OpenCV project in a subfolder "AddOn's for Windows".
My class solves problems that are faced by Windows users of OpenCV.
Please read my detailed description on StackOverflow where you also find the source code of my class.
http://stackoverflow.com/questions/24725155/opencv-tesseract-how-to-replace-libpng-libtiff-etc-with-gdi-bitmap-load-in
```
## History
##### Daniil Osokin on 2014-07-14 06:13
```
Hi, thanks for response! You can try to make a pull request (http://code.opencv.org/projects/opencv/wiki/How_to_contribute) with this (since you have the code), developers will guide you.
- Assignee set to Robby Longhorn
- Category set to highgui-images
```
| auto-transferred,priority: normal,feature,category: imgcodecs | low | Minor |
97,433,745 | opencv | opencv 3.0 - CUDA link error in opencv_core300 --> __cudaRegisterLinkedBinary unresolved symbol | Transferred from http://code.opencv.org/issues/3816
```
|| jean-claude rosichini on 2014-07-16 19:37
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: build/install
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Windows
```
## opencv 3.0 - CUDA link error in opencv_core300 --> __cudaRegisterLinkedBinary unresolved symbol
```
Hi,
Just went through a compile/link error on opencv_core module (no other module seems suffering from the same error)
My config:
use latest trunk opencv 3.0 version (as of July 16th 2014)
VS2010 64bits
CUDA 6.0 (btw same issue with 6.5)
*symptom:*
_4> Creating library C:/Users/jc/Desktop/DEVELOPMENT_LIBS/opencv/x64/lib/Debug/opencv_core300d.lib and object C:/Users/jc/Desktop/DEVELOPMENT_LIBS/opencv/x64/lib/Debug/opencv_core300d.exp
4>cuda_compile_generated_gpu_mat.cu.obj : error LNK2019: unresolved external symbol __cudaRegisterLinkedBinary_54_tmpxft_00002bd0_00000000_11_gpu_mat_compute_35_cpp1_ii_71482d89 referenced in function "void __cdecl __sti____cudaRegisterAll_54_tmpxft_00002bd0_00000000_11_gpu_mat_compute_35_cpp1_ii_71482d89(void)" (?__sti____cudaRegisterAll_54_tmpxft_00002bd0_00000000_11_gpu_mat_compute_35_cpp1_ii_71482d89@@YAXXZ)
4>C:\Users\jc\Desktop\DEVELOPMENT_LIBS\opencv\x64\bin\Debug\opencv_core300d.dll : fatal error LNK1120: 1 unresolved externals
4>
4>Build FAILED.
_
thanks
```
## History
##### Dmitry Retinskiy on 2014-09-17 07:41
```
Jean-Claude, please attach full build log.
Alexander, could you check this?
Thanks.
- Assignee set to Alexander Smorkalov
- Status changed from New to Open
- Category set to build/install
```
| bug,auto-transferred,priority: normal,category: build/install,affected: 3.4 | low | Critical |
97,433,909 | opencv | Horn & Lucas Kanade Legacy | Transferred from http://code.opencv.org/issues/3823
```
|| Juan Perez Rua on 2014-07-22 12:36
|| Priority: Low
|| Affected: None
|| Category: legacy, contrib
|| Tracker: Feature
|| Difficulty:
|| PR:
|| Platform: None / None
```
## Horn & Lucas Kanade Legacy
```
Hello,
I was just wondering if its really helpful to have Horn and Lucas-Kanade optical flow methods as legacy/obsolete as they are used even in modern fusion-based optical flow methods (like the fusionflow). The point is that these methods though old are still referents in the motion estimation state of the art. I could do some work with the c++ interface if you think it shouldn't be labeled as legacy.
Juan
```
## History
##### Dmitry Retinskiy on 2014-09-17 10:41
```
- Category set to legacy, contrib
```
| auto-transferred,feature,priority: low,category: legacy,category: contrib | low | Minor |
97,433,942 | opencv | Concurrency in cuda module fails with visual studio | Transferred from http://code.opencv.org/issues/3824
```
|| Ernest Galbrun on 2014-07-23 12:12
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: gpu (cuda)
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Windows
```
## Concurrency in cuda module fails with visual studio
```
I have had trouble using the cuda functions in a multithreaded environment. After some investigation, it appears that the issue comes from the function-scope static members used in cuda_stream.cpp
As specified here, function-scope static initialization is not thread safe:
http://msdn.microsoft.com/en-us/library/vstudio/s1sb61xd(v=vs.120).aspx
I suggest using a file-scope static variable and an excplicit initialization function since we want to defer the execution of the initialization until we need the variable.
I have implemented the changes but since it deals with core functionality of cuda opencv I would like a direct opinion about that.
```
## History
##### Dmitry Retinskiy on 2014-09-17 07:35
```
Ernest,
thanks for the suggested correction.
Please put it in the form of pull request (see http://www.code.opencv.org/projects/opencv/wiki/How_to_contribute).
Each pull request is automatically tested and then reviewed before merge.
- Assignee set to Ernest Galbrun
- Status changed from New to Open
- Category set to gpu (cuda)
```
| bug,auto-transferred,priority: normal,affected: 3.4,category: gpu/cuda (contrib) | low | Critical |
97,433,968 | opencv | Nvcc error: Only when targeting sm_20 or higher | Transferred from http://code.opencv.org/issues/3827
```
|| hoa Hong on 2014-07-24 05:11
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: gpu (cuda)
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x86 / Windows
```
## Nvcc error: Only when targeting sm_20 or higher
```
Hi,
I'm building latest opencv 3.0 from source with CUDA SDK 5.5. My graphics card is Gigabyte NVidia GeForce 9800GTX+ with 1.1 computability. When building opencv_core module, i get this error:
"*nvcc : fatal error : nvcc supports '--relocatable-device-code=true (-rdc=true)' , '--device-c (-dc)' and '--device-link (-dlink)' only when targeting sm_20 or higher*"
CMake settings:
CUDA_ARCH_BIN: 1.1 1.2 1.3 2.0 2.2(2.0) 3.0.3.5
CUDA_ARCH_PTX: 3.0
Enviroment and tools:
Visual Studio C++ 2012
Windows 7
I changed CUDA_ARCH_BIN and CUDA_ARCH_PTX but i seems that doesn't affect.
Is it possible a bug?
```
## History
##### Steven Puttemans on 2014-07-25 07:32
```
Since 3.0 is a development branch OpenCV does not guarantee it will build without problems. I experience similar GPU building problems with 3.0 compute capability, seems this is indeed an unfixed bug!
```
##### hoa Hong on 2014-07-27 15:32
```
Steven Puttemans wrote:
> Since 3.0 is a development branch OpenCV does not guarantee it will build without problems. I experience similar GPU building problems with 3.0 compute capability, seems this is indeed an unfixed bug!
Do you know how to fix it?
```
##### Steven Puttemans on 2014-07-28 07:14
```
Nope :) My GPU experience is limited to using the functions :)
```
##### Vadim Pisarevsky on 2015-04-28 11:42
```
- Category changed from core to gpu (cuda)
```
| bug,auto-transferred,priority: normal,affected: 3.4,category: gpu/cuda (contrib) | low | Critical |
97,433,993 | opencv | Build error on Windows when CUDA is disabled | Transferred from http://code.opencv.org/issues/3830
```
|| Erfan Noury on 2014-07-26 12:19
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: build/install
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Windows
```
## Build error on Windows when CUDA is disabled
```
I'm on Windows 8.1 x64 with CUDA 6.0 and VS2013 installed, but since CUDA 6.0 doesn't have compatibility with VS2013, I disabled CUDA in CMake configuration.
Problem is, when trying to build Opencv,I get link error in Visual Studio, since projects are linked against CUDA libraries.
Here is my CMake output:
<pre>
Checking for Windows Platform SDK
Checking for Visual Studio 2012
Could NOT find Matlab (missing: MATLAB_MEX_SCRIPT MATLAB_INCLUDE_DIRS MATLAB_ROOT_DIR MATLAB_LIBRARIES MATLAB_LIBRARY_DIRS MATLAB_MEXEXT MATLAB_ARCH MATLAB_BIN)
General configuration for OpenCV 3.0.0-dev =====================================
Version control: 2.4.9-4346-gda1ac35-dirty
Platform:
Host: Windows 6.2 AMD64
CMake: 2.8.12.1
CMake generator: Visual Studio 12 Win64
CMake build tool: C:/PROGRA~2/MSBuild/12.0/Bin/MSBuild.exe
MSVC: 1800
C/C++:
Built as dynamic libs?: YES
C++ Compiler: C:/Program Files (x86)/Microsoft Visual Studio 12.0/VC/bin/x86_amd64/cl.exe (ver 18.0.30501.0)
C++ flags (Release): /DWIN32 /D_WINDOWS /W4 /GR /EHa /D _CRT_SECURE_NO_DEPRECATE /D _CRT_NONSTDC_NO_DEPRECATE /D _SCL_SECURE_NO_WARNINGS /Gy /bigobj /arch:AVX /Oi /wd4251 /MD /O2 /Ob2 /D NDEBUG /Zi
C++ flags (Debug): /DWIN32 /D_WINDOWS /W4 /GR /EHa /D _CRT_SECURE_NO_DEPRECATE /D _CRT_NONSTDC_NO_DEPRECATE /D _SCL_SECURE_NO_WARNINGS /Gy /bigobj /arch:AVX /Oi /wd4251 /D_DEBUG /MDd /Zi /Ob0 /Od /RTC1
C Compiler: C:/Program Files (x86)/Microsoft Visual Studio 12.0/VC/bin/x86_amd64/cl.exe
C flags (Release): /DWIN32 /D_WINDOWS /W3 /D _CRT_SECURE_NO_DEPRECATE /D _CRT_NONSTDC_NO_DEPRECATE /D _SCL_SECURE_NO_WARNINGS /Gy /bigobj /arch:AVX /Oi /MD /O2 /Ob2 /D NDEBUG /Zi
C flags (Debug): /DWIN32 /D_WINDOWS /W3 /D _CRT_SECURE_NO_DEPRECATE /D _CRT_NONSTDC_NO_DEPRECATE /D _SCL_SECURE_NO_WARNINGS /Gy /bigobj /arch:AVX /Oi /D_DEBUG /MDd /Zi /Ob0 /Od /RTC1
Linker flags (Release): /machine:x64 /INCREMENTAL:NO /debug
Linker flags (Debug): /machine:x64 /debug /INCREMENTAL
Precompiled headers: YES
OpenCV modules:
To be built: core flann imgproc imgcodecs videoio highgui features2d calib3d ml nonfree objdetect optim photo video shape stitching superres ts videostab
Disabled: world
Disabled by dependency: -
Unavailable: androidcamera cuda cudaarithm cudabgsegm cudacodec cudafeatures2d cudafilters cudaimgproc cudalegacy cudaoptflow cudastereo cudawarping cudev java matlab viz
Windows RT support: NO
GUI:
QT: NO
Win32 UI: YES
OpenGL support: NO
VTK support: NO
Media I/O:
ZLib: build (ver 1.2.8)
JPEG: build (ver 90)
WEBP: NO
PNG: build (ver 1.5.12)
TIFF: build (ver 42 - 4.0.2)
JPEG 2000: build (ver 1.900.1)
OpenEXR: NO
Video I/O:
Video for Windows: YES
DC1394 1.x: NO
DC1394 2.x: NO
FFMPEG: YES (prebuilt binaries)
codec: YES (ver 55.18.102)
format: YES (ver 55.12.100)
util: YES (ver 52.38.100)
swscale: YES (ver 2.3.100)
gentoo-style: YES
OpenNI: NO
OpenNI PrimeSensor Modules: NO
OpenNI2: NO
PvAPI: NO
GigEVisionSDK: NO
DirectShow: YES
Media Foundation: YES
XIMEA: NO
Intel PerC: NO
Other third-party libraries:
Use IPP: NO
Use IPP Async: NO
Use Eigen: YES (ver 3.2.1)
Use TBB: YES (ver 4.2 interface 7005)
Use OpenMP: NO
Use GCD NO
Use Concurrency NO
Use C=: NO
Use Cuda: NO
Use OpenCL: NO
Python:
Interpreter: C:/Anaconda/python.exe (ver 2.7.6)
Java:
ant: NO
JNI: C:/Program Files/Java/jdk1.8.0_05/include C:/Program Files/Java/jdk1.8.0_05/include/win32 C:/Program Files/Java/jdk1.8.0_05/include
Java tests: NO
Matlab:
mex: NO
Tests and samples:
Tests: NO
Performance tests: YES
C/C++ Examples: NO
</pre>
And this is the error inside Visual Studio:
<pre>
error LNK1104: cannot open file 'opencv_cudev.lib' opencv_core
</pre>
```
## History
| bug,auto-transferred,priority: normal,category: build/install,affected: 3.4 | low | Critical |
97,434,071 | opencv | XIMEA Image format incorrect after DOWNSAMPLING | Transferred from http://code.opencv.org/issues/3836
```
|| Jean-Yves Deschenes on 2014-07-31 15:12
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: highgui-camera
|| Tracker: Bug
|| Difficulty: Easy
|| PR:
|| Platform: x64 / Windows
```
## XIMEA Image format incorrect after DOWNSAMPLING
```
If I open a XIMA camera and set the downsampling, the retrieved capture frames will have the wrong dimensions.
<pre>
// Open the camera:
cv::VideoCapture cam;
cv::Mat bufferFrame;
int columns = 0;
cam.open(inputName);
cam.read(bufferFrame);
columns = bufferFrame.columns; // _Columns value is 2592_
cam.set(CV_CAP_PROP_XI_DATA_FORMAT, XI_RGB24 );
cam.set(CV_CAP_PROP_XI_DOWNSAMPLING, ximeaDefaultBinningMode);
// In the thread loop
cam.read(bufferFrame);
columns = bufferFrame.columns; // _Columns value is still 2592_
</pre>
Now, that seems to come from the fact that, in cap_ximea.cpp,
<pre>
void CvCaptureCAM_XIMEA::resetCvImage() // Called from CvCaptureCAM_XIMEA::retrieveFrame()
{
...
if( (int)image.width != width || (int)image.height != height || image.frm != (XI_IMG_FORMAT)format)
...
}
</pre>
the validation checks for the identity of the _image_ dimensions (which are the same, since they were updated by the internal CvCaptureCAM_XIMEA::grabFrame() call), but does not check for the identity of the _frame_ dimensions (which have not been changed, and need to).
An easy fix probably would be to add a check of the frame.cols and frame.rows in that condition line.
```
## History
##### Dmitry Retinskiy on 2014-09-18 05:49
```
Jean-Yves, thanks for the report.
Could you put your suggested correction in the form of pull request?
(see http://www.code.opencv.org/projects/opencv/wiki/How_to_contribute for details).
Thanks.
- Target version deleted (2.4.10)
- Description changed from If I open a XIMA camera and set the
downsampling, the retrieved capture frame... to If I open a XIMA
camera and set the downsampling, the retrieved capture frame... More
- Assignee set to Jean-Yves Deschenes
- Status changed from New to Open
```
| bug,auto-transferred,priority: normal,affected: 2.4,category: videoio(camera) | low | Critical |
97,434,102 | opencv | calcOpticalFlowPyrLK() returns points outside the image | Transferred from http://code.opencv.org/issues/3837
```
|| Philipp Hasper on 2014-07-31 17:40
|| Priority: Normal
|| Affected: branch '2.4' (2.4-dev)
|| Category: imgproc, video
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: Any / Any
```
## calcOpticalFlowPyrLK() returns points outside the image
```
calcOpticalFlowPyrLK() returns points outside the image and marks them as valid (status[i]==true). So far I only observed points with coordinates slightly bigger than the boundaries but no negative values.
This issue is easily reproducible since it occurs in nearly every call of the function.
Just run the attached code (this is the lkdemo.cpp example with one simple assertion added in line 109):
I could not track down the bug maybe it happens in lkpyramid.cpp, function @LKTrackerInvoker::operator()@, line 428: @nextPts[ptidx] = nextPt + halfWin;@ Here we just add a window width and this may lead to a point outside the image.
```
## History
##### Philipp Hasper on 2014-07-31 18:08
```
I have to correct: I also experienced negative values.
```
##### Dmitry Retinskiy on 2014-09-16 07:03
```
Vadim, please check the problem.
Thanks.
- Priority changed from High to Normal
- Assignee set to Vadim Pisarevsky
- Status changed from New to Open
- Category set to imgproc, video
```
| bug,auto-transferred,priority: normal,category: imgproc,category: video,affected: 2.4 | low | Critical |
97,434,147 | opencv | Reading Image without Meta-Data fails | Transferred from http://code.opencv.org/issues/3838
```
|| M B on 2014-08-04 12:29
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: highgui-images
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x86 / Windows
```
## Reading Image without Meta-Data fails
```
* Bug concering imread -> imread does not read image without metadata
* fixed by editing image meta data (e.g. setting a title or cameratype)
* testimages:
not working: https://docs.google.com/file/d/0B5AU_enXY31NQlJ2eVU2WmdwRVk/edit
and working: https://docs.google.com/file/d/0B5AU_enXY31NbWxFSmJnbWhuZ2s/edit
* Asked @Q&A 4 weeks ago without an answer: http://answers.opencv.org/question/35944/imread-without-exif-title-results-in-invalid-mat/
+Long report:+
The error occurs when I try to load an 16bit tiff grayscale image without a meta data set in its details created by a camera software.
It can be viewed with ImageJ without problem.
When I try to load it with imread it results in an empty matrix, which has the wrong type, as well as one wrong dimension.
It should be UINT16, 1x768x768, but the debugger tells me it is UINT8, 1x768x0.
Not working image:
https://docs.google.com/file/d/0B5AU_enXY31NQlJ2eVU2WmdwRVk/edit
If I open the preferences of the image and change the title in details to anything and press ok, I can open it and its read perfectly.
Same image with title changed:
https://docs.google.com/file/d/0B5AU_enXY31NbWxFSmJnbWhuZ2s/edit
I use ImageWatch which tells me that my matrix has no dimensions when I try to load the first sample image.
The images are uncompressed 16bit tiff files, generated by a 12bit camera.
If I open and save the image with ImageJ and set a certain display range (eg 16bit) and save it, the described "trick" will not help anymore.
+My setup:+
Win 8.1 (64bit)
C++ VS2013 (32bit)
OpenCv 2.4.9
+My test code:+
<pre>
#include "stdafx.h"
#include <opencv2/core/core.hpp>
#include <opencv2/highgui/highgui.hpp>
using namespace cv;
using namespace std;
int _tmain(int argc, _TCHAR* argv[])
{
string path = "C:\\Users\\...\\Pictures\\testimage.tiff";
Mat A;
A = cv::imread(path, CV_LOAD_IMAGE_GRAYSCALE);
namedWindow("A", CV_WINDOW_NORMAL);
resizeWindow("A", A.cols, A.rows);
imshow("A", A);
waitKey(0);
return 0;
}
</pre>
```
## History
##### Dmitry Retinskiy on 2014-09-18 05:53
```
Alexander, could you check this?
Thanks.
- Assignee set to Alexander Smorkalov
- Status changed from New to Open
```
| auto-transferred,feature,priority: low,affected: 2.4,category: imgcodecs | low | Critical |
97,434,165 | opencv | Add support for broadcasting arithmetical and logic operands | Transferred from http://code.opencv.org/issues/3839
```
|| Goncalo Lopes on 2014-08-04 13:37
|| Priority: Normal
|| Affected: None
|| Category: core
|| Tracker: Feature
|| Difficulty:
|| PR:
|| Platform: None / None
```
## Add support for broadcasting arithmetical and logic operands
```
This allows for easy and efficient vectorization for basic arithmetic and logic, similar to NumPy:
http://docs.scipy.org/doc/numpy/user/basics.broadcasting.html
This is possible when using OpenCV in Python because of the NumPy interface, but it would be very desirable to be able to do this in C++ as well.
Also, it would be nice to be able to do this without resorting to intermediate buffer allocations and the use of cv::repeat (again, identical to NumPy broadcasting).
```
## History
| auto-transferred,priority: normal,feature,category: core | low | Minor |
97,434,280 | opencv | Cascadeclassifier_gpu with cv::gpu::stream support | Transferred from http://code.opencv.org/issues/3844
```
|| T Abdullah on 2014-08-06 09:46
|| Priority: Normal
|| Affected: None
|| Category: objdetect
|| Tracker: Feature
|| Difficulty:
|| PR:
|| Platform: None / None
```
## Cascadeclassifier_gpu with cv::gpu::stream support
```
Dear all,
I wanted to detect objects from video streams using cascadeclassifier_gpu.detectMultiScale(...) with multiple video streams in parallel. I could execute two parallel threads. The program crashed with more than two threads (See more crash details in [[http://code.opencv.org/issues/3807]] in OpenCV DevZone). I hope to have a solution for it.
I would like to suggest cv::gpu::stream support for cascadeclassifier_gpu. As this support will enable to process multiple video files on GPU in multiple streams.
Regards,
Tariq
```
## History
##### Dmitry Retinskiy on 2014-09-17 10:31
```
- Category set to objdetect
```
| auto-transferred,priority: normal,feature,category: objdetect | low | Critical |
97,434,311 | opencv | More explicit exception message when OpenGL calls made without an active context | Transferred from http://code.opencv.org/issues/3845
```
|| Musharraf Azeem on 2014-08-06 13:35
|| Priority: Normal
|| Affected: None
|| Category: infrastructure
|| Tracker: Feature
|| Difficulty:
|| PR:
|| Platform: None / None
```
## More explicit exception message when OpenGL calls made without an active context
```
Functions from ogl:: are called that require an OpenGL context to be active on the current thread. When they are called without an active context, then a error with the message "Can't load OpenGL extension [ext_name]" is thrown.
The above error could be called whenever CV_GL_GET_PROC_ADDRESS fails to load an extension, which could be for a variety of reasons
In the scenario that OpenGL calls were attempted without an active OpenGL context, then this should explicitly be stated in the error message thrown. This will be especially helpful in debugging a multithreaded application.
```
## History
##### Dmitry Retinskiy on 2014-09-17 10:30
```
Musharraf, are you ready to implement your suggestion?
(see http://www.code.opencv.org/projects/opencv/wiki/How_to_contribute for details)
- Assignee set to Musharraf Azeem
- Category set to infrastructure
```
| auto-transferred,priority: normal,feature,category: infrastructure | low | Critical |
97,434,392 | opencv | hogdescriptor::detectmultiscale to output feature vectors for detections | Transferred from http://code.opencv.org/issues/3848
```
|| Kevin Shi on 2014-08-06 16:41
|| Priority: High
|| Affected: None
|| Category: objdetect
|| Tracker: Feature
|| Difficulty:
|| PR:
|| Platform: None / None
```
## hogdescriptor::detectmultiscale to output feature vectors for detections
```
detectMultiScale currently just outputs a vector<Rect> with no additional information. To get the feature vectors for each detection, I have to crop the image to each rect and compute the features again, which is extremely time consuming and slowing down my code by several factors.
Since these features are already computed in detectMultiScale, it should be possible to pass an additional argument (vector<vector<float> >& features) to detectMultiScale and have the detector write the feature vectors for each detection to this argument.
```
## History
##### Steven Puttemans on 2014-08-07 11:10
```
I like the idea, but it should be implemented in the Algorithm class so that each descriptor detector can benefit from this.
You might be up for the job yourself and try it out? Making a pull request is quite straightforward and increase the chance your feature gets in.
```
##### Dmitry Retinskiy on 2014-09-17 10:27
```
- Target version changed from 3.0-beta to 3.0
- Category set to objdetect
```
##### Deanna Hood on 2015-04-19 20:24
```
I'm interested in working on this. I am planning to overload each of the three existing detectMultiScale methods to have a version which provides the additional argument. Please advise if there is a preferred alternative to having 6 versions of the detectMultiScale function call.
```
##### Steven Puttemans on 2015-04-20 14:23
```
Deanna Hood wrote:
> I'm interested in working on this. I am planning to overload each of the three existing detectMultiScale methods to have a version which provides the additional argument. Please advise if there is a preferred alternative to having 6 versions of the detectMultiScale function call.
Since we are forced not to break backwards compatibility I think that overload functions with an extra parameter are the best way to go! If you are up to this in a pull request, then please also update the corresponding documentation so that people are aware of the new functionality.
- Status changed from New to Open
```
##### Deanna Hood on 2015-05-09 17:11
```
Just updating that I haven't gotten around to this, so that others know it's not assigned
```
| auto-transferred,feature,category: objdetect | low | Major |
97,434,449 | opencv | Public parameters of class not accessible | Transferred from http://code.opencv.org/issues/3849
```
|| Steven Puttemans on 2014-08-07 13:33
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: legacy, contrib
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Linux
```
## Public parameters of class not accessible
```
When looking at this Q&A topic (http://answers.opencv.org/question/38851/how-to-create-a-lbp-histogram-using-opencv/) I wondered if it couldnt be easier just to publicly expose the histograms of the LBPH Face detector which is in the contrib module.
I found out that in facerec.cpp there are the following commands in the declaration of the class
<pre>
// Getter functions.
int neighbors() const { return _neighbors; }
int radius() const { return _radius; }
int grid_x() const { return _grid_x; }
int grid_y() const { return _grid_y; }
</pre>
What lead me to believe since they are declared public that they would be accesible and I could do the same for the histograms.
So changing this to:
<pre>
// Getter functions.
int neighbors() const { return _neighbors; }
int radius() const { return _radius; }
int grid_x() const { return _grid_x; }
int grid_y() const { return _grid_y; }
vector<Mat> histograms() const { return _histograms; }
</pre>
Worked perfectly for building, but then when using the function for a code snippet.
<pre>
#include "opencv2/opencv.hpp"
using namespace std;
using namespace cv;
int main()
{
// holds images and labels
vector<Mat> images;
vector<int> labels;
// images for first person
images.push_back(imread("/data/temp/1.png", CV_LOAD_IMAGE_GRAYSCALE)); labels.push_back(0);
images.push_back(imread("/data/temp/1b.png", CV_LOAD_IMAGE_GRAYSCALE)); labels.push_back(0);
// images for second person
images.push_back(imread("/data/temp/2.png", CV_LOAD_IMAGE_GRAYSCALE)); labels.push_back(1);
images.push_back(imread("/data/temp/2b.png", CV_LOAD_IMAGE_GRAYSCALE)); labels.push_back(1);
Ptr<FaceRecognizer> model = createLBPHFaceRecognizer();
model->train(images, labels);
int grid_x = model->grid_x();
vector<Mat> histograms = model->histograms();
return 0;
}
</pre>
It seems that there is an error at compile time stating that both functions are no members of FaceRecognizer.
<pre>
||=== Build: Debug in test_LBP_hist (compiler: GNU GCC Compiler) ===|
/home/spu/Projects/code_blocks/test_LBP_hist/main.cpp||In function ‘int main()’:|
/home/spu/Projects/code_blocks/test_LBP_hist/main.cpp|22|error: ‘class cv::FaceRecognizer’ has no member named ‘grid_x’|
/home/spu/Projects/code_blocks/test_LBP_hist/main.cpp|23|error: ‘class cv::FaceRecognizer’ has no member named ‘histograms’|
/home/spu/Projects/code_blocks/test_LBP_hist/main.cpp|22|warning: unused variable ‘grid_x’ [-Wunused-variable]|
||=== Build failed: 2 error(s), 1 warning(s) (0 minute(s), 0 second(s)) ===|
</pre>
This is wrong right? Since model should know that it is internally a LBPH element so those functions should be available?
```
## History
| bug,auto-transferred,priority: normal,category: legacy,category: contrib,affected: 2.4 | low | Critical |
97,434,531 | opencv | Dependency on Python version and default encoding | Transferred from http://code.opencv.org/issues/3854
```
|| Philippe FOUBERT on 2014-08-10 16:40
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: matlab bindings
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Windows
```
## Dependency on Python version and default encoding
```
h1. Dependency on Python version
I have Python 2.7.8 (the latest Python 2 release version) and Python 3.4.1 (the latest Python 3 release version) installed on my computer.
The "@modules\matlab\CMakeLists.txt@" file is using the "@${PYTHON_DEFAULT_EXECUTABLE}@" variable (which is initialized with "@${PYTHON2_EXECUTABLE}@" in "@cmake\OpenCVDetectPython.cmake@") to launch the scripts "@gen_matlab.py@" and "@cvmex.py@" which are using the *"@lstrip_blocks@"* parameter which seems *not available in Python 2.7.8* :
> @jtemplate = Environment(loader=FileSystemLoader(template_dir), trim_blocks=True, lstrip_blocks=True)@
If the "@${PYTHON_DEFAULT_EXECUTABLE}@" variable is replaced by the "@${PYTHON3_EXECUTABLE}@" variable, the error disappears.
h1. Dependency on unicode
I am working on a 64 bits OS. On such an OS, the default encoding used by Python is unicode.
=> in the "@modules\java\generator\rst_parser.py@" script, I have had to add the encoding to the open function:
> @df = open(doc, "rt" , encoding="utf_8")@
a parameter which is not compatible with Python2...
h1. Default Python executable
Since the default Python executable is not only used in the Matlab binding module, i have chosen to write an issue instead of a PR...
```
## History
| bug,auto-transferred,priority: normal,affected: 3.4,category: matlab bindings | low | Critical |
97,434,615 | opencv | Expectation maximization covariance not accepted when feeding back previous results in python | Transferred from http://code.opencv.org/issues/3862
```
|| Santiago Peñate Vera on 2014-08-14 15:00
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: python bindings
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Windows
```
## Expectation maximization covariance not accepted when feeding back previous results in python
```
I want to train a picture, get the means, covariances, etc.. and feed them back into another train of the same picture but a bit displaced.
Apparently there is no possible way to re-feed in the covariances using EM::trainE
See the provided example at the file Test_EM.py
```
## History
| bug,auto-transferred,priority: normal,category: python bindings,affected: 2.4 | low | Critical |
97,434,731 | opencv | DynamicAdaptedFeatureDetector for image sequences | Transferred from http://code.opencv.org/issues/3873
```
|| Felix Endres on 2014-08-19 13:25
|| Priority: Normal
|| Affected: None
|| Category: features2d
|| Tracker: Feature
|| Difficulty: Medium
|| PR:
|| Platform: None / None
```
## DynamicAdaptedFeatureDetector for image sequences
```
Hi all,
I propose a modification (or a new variant of) the @DynamicAdaptedFeatureDetector@ and @AdjusterAdapter@.
I use the opencv visual features in a mapping application.
I noticed that the number of keypoints is crucial for the performance of the motion estimates.
If there are too few keypoints, too few matches are found. If there are too many, the detection,
extraction and matching time is high and more ambiguities arise during matching.
I once tried to use the above classes to control the amount of keypoints, but found them to create
a big performance penalty. I recently found the reasons:
* The change rate of the thresholds is only 10% in each iteration
* The found threshold is not stored for the next call to detect(...)
This results in many iterations that are repeated for each image in the stream.
I rewrote the mentioned classes to overcome these issues and added further small improvements.
Also, I patched ORB to expose the threshold used for the FAST keypoints, such that it is adjustable.
If someone would guide me and discuss the changes with me, I would be happy to contribute them back.
Felix
```
## History
##### Steven Puttemans on 2014-08-20 09:31
```
First of all thanks for the changes! In order to speed up the process of adapting everything and making the changes in source code it would be nice if you created a pull request.
Do keep in mind that extra functionality goes to the 3.0 dev branch and should have documentation and performance tests added, in which you can be helped when the pull request is up.
Could you try creating one? Guidelines can be found here: http://code.opencv.org/projects/opencv/wiki/How_to_contribute
- Status changed from New to Open
```
##### Steven Puttemans on 2014-08-20 09:32
```
- Target version set to 3.0
- Difficulty set to Medium
```
##### Philipp Hasper on 2014-10-06 10:39
```
+1 Nice contribution!
```
##### Felix Endres on 2014-11-28 11:04
```
This is a bit embarrassing, but I cannot find the original AdjusterAdapter and DynamicAdaptedFeatureDetector in the OpenCV repositories in the master branch. The file modules/features2d/dynamic.cpp is more or less empty. After a while I found out that they should have moved to opencv_contrib (according to the commit message of 31df47b6eaeb753a77f4d0cc3f4225c5c40682f6: "moved nonfree and a part of features2d to opencv_contrib/xfeatures2d"). However, there I cannot find anything. Do you have any pointers? The proposed code is actually independent of that, but if the orignal "dynamic" parts are still there, it definitely belongs to the same module/file.
Otherwise, where should I place it?
```
| auto-transferred,priority: normal,feature,category: features2d | low | Major |
97,434,806 | opencv | Removed FilterEngine from public API | Transferred from http://code.opencv.org/issues/3879
```
|| Philippe FOUBERT on 2014-08-23 10:22
|| Priority: Normal
|| Affected: None
|| Category: imgproc, video
|| Tracker: Feature
|| Difficulty: Easy
|| PR:
|| Platform: None / None
```
## Removed FilterEngine from public API
```
I think the removal of *FilterEngine* from the public API is not a good idea.
If we need to perform the same filter cyclically, it is very usefull (*real-time efficiency*) to be able to allocate and initialize the filter only once.
For example:
> @cv::Ptr<cv::FilterEngine> pFilter;
> // At init:
> pFilter = createBoxFilter(CV_32F, CV_32F, cv::Size(15, 15), cv::Point(-1, -1), true, cv::BORDER_DEFAULT);
> // Cyclically:
> pFilter->apply(src, dst);
> @
instead of calling cyclically:
> @cv::boxFilter(src, dst, CV_32F, cv::Size(15, 15), cv::Point(-1, -1), true, cv::BORDER_DEFAULT);
> @
```
## History
| auto-transferred,priority: normal,feature,category: imgproc,category: video | low | Minor |
97,434,830 | opencv | Contrib: fcw_detect and fcw_train have wrong dependencies | Transferred from http://code.opencv.org/issues/3881
```
|| Philipp Hasper on 2014-08-25 12:32
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: build/install
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: Any / Any
```
## Contrib: fcw_detect and fcw_train have wrong dependencies
```
There seems to be an error in the CMake chain: the opencv_fcw_detect and opencv_fcw_train projects which are generated when we integrate the contrib module have dependencies to
> opencv_core.lib
> opencv_imgcodecs.lib
> opencv_videoio.lib
> opencv_highgui.lib
even if the approach using opencv_world300.lib was selected and thus the libraries mentioned above are not created.
The result is a linker error.
```
## History
##### Dmitry Retinskiy on 2014-09-16 11:27
```
Roman, could you check this?
Thanks.
- Assignee set to Roman Donchenko
- Status changed from New to Open
- Category set to build/install
```
| bug,auto-transferred,priority: normal,category: build/install,affected: 3.4 | low | Critical |
97,434,854 | opencv | MINGW 64bit detection is wrong for the most popular mingw distribution | Transferred from http://code.opencv.org/issues/3884
```
|| Yaron Keren on 2014-08-27 16:58
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: build/install
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x86 / Windows
```
## MINGW 64bit detection is wrong for the most popular mingw distribution
```
OpenCVConfig.cmake:89 says
if(CMAKE_OPENCV_GCC_TARGET_MACHINE MATCHES "64")
set(MINGW64 1)
set(OpenCV_ARCH x64)
else()
set(OpenCV_ARCH x86)
endif()
which is wrong for the most popular mingw distribution, mingw-w64 whose machinedump is "i686-w64-mingw32" even on 32 bit targets. This logic will always identify it as 64 bit.
Moreover, this logic is incompatible with OpenCVDetectCXXCompiler.cmake:96 which is supposed to do the same function:
if(CMAKE_OPENCV_GCC_TARGET_MACHINE MATCHES "amd64|x86_64|AMD64")
set(MINGW64 1)
endif()
this logic will never happen.
```
## History
##### Dmitry Retinskiy on 2014-09-16 11:17
```
Roman, could you check this?
Thanks.
- Assignee set to Roman Donchenko
- Status changed from New to Open
- Category set to build/install
```
| bug,auto-transferred,priority: normal,category: build/install,affected: 3.4 | low | Critical |
97,435,309 | opencv | Keypoint detection not deterministic (java opencv) | Transferred from http://code.opencv.org/issues/3903
```
|| Moe Doe on 2014-09-12 08:12
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: features2d
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Windows
```
## Keypoint detection not deterministic (java opencv)
```
As described in these stackoverflow post (http://stackoverflow.com/questions/25752497/keypoint-matching-just-works-two-times-java-opencv), a simple surf detection gives me different results if i run the detection several times in one application lifetime.
If I restart the application it starts again detecting two times correct and afterwards it results in different keypoints...
```
## History
##### Dmitry Retinskiy on 2014-09-16 07:16
```
Maria, could you check the problem?
Thanks.
- Assignee set to Maria Dimashova
- Status changed from New to Open
- Category set to features2d
```
| bug,auto-transferred,priority: normal,affected: 2.4,category: features2d | low | Critical |
97,435,347 | opencv | Highgui window is not scaled properly when displaying small images | Transferred from http://code.opencv.org/issues/3906
```
|| Khue Le on 2014-09-14 15:48
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: highgui-gui
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Linux
```
## Highgui window is not scaled properly when displaying small images
```
In OpenCV 3.0 (I did not try other versions), when displaying small images, even with the `WINDOW_AUTOSIZE` flag, the window is not scaled properly.
<pre>
#include <opencv2/core.hpp>
#include <opencv2/imgcodecs.hpp>
#include <opencv2/highgui.hpp>
#include <iostream>
using namespace cv;
using namespace std;
int main( int argc, char** argv )
{
if( argc != 2)
{
cout <<" Usage: display_image ImageToLoadAndDisplay" << endl;
return -1;
}
Mat image;
image = imread(argv[1], IMREAD_COLOR); // Read the file
if( image.empty() ) // Check for invalid input
{
cout << "Could not open or find the image" << std::endl ;
return -1;
}
namedWindow( "Display window", WINDOW_AUTOSIZE ); // Create a window for display.
imshow( "Display window", image ); // Show our image inside it.
while (waitKey(0) != 27); // Wait for ESC key
return 0;
}
</pre>
Test image (100x100):
!http://i.stack.imgur.com/6XB1v.png!
Results:
!http://i.stack.imgur.com/LZOU7.png!
When moving the mouse to the displayed image, I got this:
!http://i.stack.imgur.com/Fo9kx.png!
```
## History
##### Dmitry Retinskiy on 2014-09-16 07:05
```
- Description changed from In OpenCV 3.0 (I did not try other
versions), when displaying small images, e... to In OpenCV 3.0 (I
did not try other versions), when displaying small images, e... More
```
##### Dmitry Retinskiy on 2014-09-16 07:12
```
Alexander, please check the problem.
Thanks.
- Target version set to 3.0
- Assignee set to Alexander Smorkalov
- Status changed from New to Open
- Category set to highgui-gui
```
##### Theodore Tsesmelis on 2015-03-08 18:01
```
I can confirm the above bug. I noticed that it is related to the integration with the Qt library. If I compile opencv v.3.0.0 (master branch from github) with Qt4 I do not get this bug, on the other hand if I compile it with Qt5 then it appears. Moreover, on the command line when I am running a project which has to imshow an image I get the following message: *QSettings::value: Empty key passed* twice
```
##### Steven Puttemans on 2015-03-09 08:25
```
Theodore Tsesmelis wrote:
> I can confirm the above bug. Moreover, I noticed that it is related to the integration with the Qt library. If I compile opencv v.3.0.0 (master branch from github) with Qt4 I do not get this bug, on the other hand if I compile it with Qt5 then it appears. Moreover, on the command line when I am running a project which has to imshow an image I get the following message: *QSettings::value: Empty key passed* twice
I can concur that this indeed only happens with Qt5 installations. Got same problem here. My solution was to ignore Qt supported builds for the moment, not really need that extra functionality.
```
| bug,auto-transferred,priority: normal,category: highgui-gui,affected: 3.4 | low | Critical |
97,435,384 | opencv | videoio compilation error in cap_msmf.cpp | Transferred from http://code.opencv.org/issues/3907
```
|| Christoph Pacher on 2014-09-17 15:43
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: highgui-video
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x86 / Windows
```
## videoio compilation error in cap_msmf.cpp
```
at line 780 MF_MT_FRAME_RATE_RANGE_MAX and line 788 MF_MT_FRAME_RATE_RANGE_MIN are undefined
```
## History
##### Christoph Pacher on 2014-09-17 15:49
```
the rest of the errors
Error 5 error C2065: 'MF_MT_AAC_PAYLOAD_TYPE' : undeclared identifier \modules\videoio\src\cap_msmf.cpp 839
Error 6 error C2065: 'MF_MT_AAC_AUDIO_PROFILE_LEVEL_INDICATION' : undeclared identifier \modules\videoio\src\cap_msmf.cpp 840
Error 7 error C2065: 'MF_MT_FRAME_RATE_RANGE_MAX' : undeclared identifier \modules\videoio\src\cap_msmf.cpp 843
Error 8 error C2065: 'MF_MT_FRAME_RATE_RANGE_MIN' : undeclared identifier \modules\videoio\src\cap_msmf.cpp 844
Error 9 error C2065: 'MF_MT_ARBITRARY_HEADER' : undeclared identifier \modules\videoio\src\cap_msmf.cpp 879
Error 10 error C2065: 'MF_MT_ARBITRARY_FORMAT' : undeclared identifier \modules\videoio\src\cap_msmf.cpp 880
Error 11 error C2065: 'MF_MT_IMAGE_LOSS_TOLERANT' : undeclared identifier \modules\videoio\src\cap_msmf.cpp 881
Error 12 error C2065: 'MF_MT_MPEG4_SAMPLE_DESCRIPTION' : undeclared identifier \modules\videoio\src\cap_msmf.cpp 882
Error 13 error C2065: 'MF_MT_MPEG4_CURRENT_SAMPLE_ENTRY' : undeclared identifier \modules\videoio\src\cap_msmf.cpp 883
Error 14 error C2065: 'MF_MT_ORIGINAL_4CC' : undeclared identifier \modules\videoio\src\cap_msmf.cpp 884
Error 15 error C2065: 'MF_MT_ORIGINAL_WAVE_FORMAT_TAG' : undeclared identifier \modules\videoio\src\cap_msmf.cpp 885
Error 16 error C2065: 'MF_SAMPLEGRABBERSINK_IGNORE_CLOCK' : undeclared identifier \modules\videoio\src\cap_msmf.cpp 1257
Error 17 error C2065: 'MF_DEVSOURCE_ATTRIBUTE_SOURCE_TYPE' : undeclared identifier \modules\videoio\src\cap_msmf.cpp 1729
Error 18 error C2065: 'MF_DEVSOURCE_ATTRIBUTE_SOURCE_TYPE_VIDCAP_GUID' : undeclared identifier \modules\videoio\src\cap_msmf.cpp 1731
Error 19 error C2065: 'MF_DEVSOURCE_ATTRIBUTE_FRIENDLY_NAME' : undeclared identifier \modules\videoio\src\cap_msmf.cpp 1954
Error 20 error C3861: 'MFEnumDeviceSources': identifier not found \modules\videoio\src\cap_msmf.cpp 2037
Error 21 error C2065: 'MF_DEVSOURCE_ATTRIBUTE_FRIENDLY_NAME' : undeclared identifier \modules\videoio\src\cap_msmf.cpp 2045
Error 22 error C2065: 'MF_DEVSOURCE_ATTRIBUTE_SOURCE_TYPE' : undeclared identifier \modules\videoio\src\cap_msmf.cpp 2146
Error 23 error C2065: 'MF_DEVSOURCE_ATTRIBUTE_SOURCE_TYPE_VIDCAP_GUID' : undeclared identifier \modules\videoio\src\cap_msmf.cpp 2148
Error 24 error C3861: 'MFEnumDeviceSources': identifier not found \modules\videoio\src\cap_msmf.cpp 2681
Error 25 error C2065: 'MFVideoFormat_H263' : undeclared identifier \modules\videoio\src\cap_msmf.cpp 3878
Error 26 error C2504: 'IMFSinkWriter' : base class undefined C:\Program Files (x86)\Windows Kits\8.1\Include\winrt\wrl\client.h 57
Error 27 error C2039: 'Finalize' : is not a member of 'Microsoft::WRL::Details::RemoveIUnknownBase<T>' \modules\videoio\src\cap_msmf.cpp 3947
Error 28 error C2065: 'MF_READWRITE_ENABLE_HARDWARE_TRANSFORMS' : undeclared identifier \modules\videoio\src\cap_msmf.cpp 3994
Error 29 error C3861: 'MFCreateSinkWriterFromURL': identifier not found \modules\videoio\src\cap_msmf.cpp 3999
Error 30 error C2039: 'AddStream' : is not a member of 'Microsoft::WRL::Details::RemoveIUnknownBase<T>' \modules\videoio\src\cap_msmf.cpp 4039
Error 31 error C2039: 'SetInputMediaType' : is not a member of 'Microsoft::WRL::Details::RemoveIUnknownBase<T>' \modules\videoio\src\cap_msmf.cpp 4074
Error 32 error C2039: 'BeginWriting' : is not a member of 'Microsoft::WRL::Details::RemoveIUnknownBase<T>' \modules\videoio\src\cap_msmf.cpp 4080
Error 33 error C2039: 'WriteSample' : is not a member of 'Microsoft::WRL::Details::RemoveIUnknownBase<T>' \modules\videoio\src\cap_msmf.cpp 4162
- Assignee deleted (Vadim Pisarevsky)
```
##### Artur Wieczorek on 2014-09-30 07:31
```
Could you please tell what compiler/toolset did you use? Was it MinGW?
Under what OS did you compile the library and what was the target OS? Win RT? Win 8?
```
##### Vladislav Vinogradov on 2014-09-30 12:22
```
Hello Christoph,
Please provide more info about the issue (OS, compiler).
- Assignee set to Christoph Pacher
- Status changed from New to Incomplete
```
##### Christoph Rackwitz on 2014-10-31 10:54
```
I get the same error. win7 x64, opencv-2.4.9, cmade VS 2012 project.
in "c:\Program Files (x86)\Windows Kits\8.0\Include\um\mfapi.h" stuff like MF_MT_FRAME_SIZE is defined, but not stuff like MF_MT_FRAME_RATE_RANGE_MAX.
what windows kit do you recommend?
```
##### Artur Wieczorek on 2014-11-17 23:12
```
> in "c:\Program Files (x86)\Windows Kits\8.0\Include\um\mfapi.h" stuff like MF_MT_FRAME_SIZE is defined, but not stuff like MF_MT_FRAME_RATE_RANGE_MAX.
>
This is strange. It seems that MF_MT_FRAME_RATE_RANGE_MAX is defined in mfapi.h taken from both SDK 7 and 8. Maybe you have some non-standard version of this header file? What is the size of your mfapi.h file? On my machine this is 78,019 bytes for SDK 7 and 107,631 bytes for SDK 8.
```
##### Vadim Pisarevsky on 2015-04-28 12:00
```
- Category changed from imgproc, video to highgui-video
```
| bug,auto-transferred,priority: low,category: videoio,category: build/install,affected: 3.4 | low | Critical |
97,435,437 | opencv | VideoCapture::read(Mat) ignores CV_CAP_PROP_CONVERT_RGB flag for H264 YUV4:2:0 file | Transferred from http://code.opencv.org/issues/3909
```
|| Vit Shiryaev on 2014-09-18 22:31
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: highgui-video
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Windows
```
## VideoCapture::read(Mat) ignores CV_CAP_PROP_CONVERT_RGB flag for H264 YUV4:2:0 file
```
When I am trying to get gray-scale frame component from H264 YVU 4:2:0 video ( Y-component from YVU 4:2:0 ) a lot of time is wasting for useless default YUV->RGB conversion while VideoCapture::read(Mat) perfoms. And for some reasone
@VideoCapture::cap.set(CV_CAP_PROP_CONVERT_RGB, false)@
gives no effect (from openCV-docs "CV_CAP_PROP_CONVERT_RGB Boolean flags indicating whether images should be converted to RGB."), as I can see from splitting:
@Mat ch[3]; split(frame,ch); imshow("it must be U-channel", ch[1]);@
frames are still in RGB, not YUV, i.e. VideoCapture::read(Mat) *ignores* CV_CAP_PROP_CONVERT_RGB flag and yuv420_bgr24_MMXEXT from opencv_ffmpeg249.dll still takes much time as I can see from AMD CodeXL profiler.
```
## History
| auto-transferred,feature,priority: low,category: videoio,affected: 2.4,incomplete | low | Critical |
97,435,479 | opencv | Intel® RealSense™ SDK (former Intel® Perceptual Computing SDK) support | Transferred from http://code.opencv.org/issues/3917
```
|| Eric Guleksen on 2014-09-25 03:04
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: build/install
|| Tracker: Bug
|| Difficulty: Medium
|| PR:
|| Platform: Any / Any
```
## Intel® RealSense™ SDK (former Intel® Perceptual Computing SDK) support
```
Prior versions of OpenCV supported the former Intel Perceptual Computing SDK natively after switching flags in the CMake file while rebuilding OpenCV. Due to a naming change and reorganization of the SDK, OpenCV support no longer is functional.
Specific Errors when attempting to build OpenCV with this new SDK:
The Environment variable PCSDK_DIR is no longer set by the SDK, it has been renamed to RSSDK_DIR, so all files which look for this environment variable will no longer function
Specifically, the pxcsmartptr.h, and pxcscheduler.h files are missing. Different implementation is needed to ensure successful build.
```
## History
##### Vladislav Vinogradov on 2014-09-30 10:46
```
Hello Eric,
Thanks for your report.
Could you fix the issue on your side and submit a pull request (http://code.opencv.org/projects/opencv/wiki/How_to_contribute)?
Your contribution will be highly appreciated!
- Assignee set to Eric Guleksen
- Status changed from New to Open
- Category set to build/install
```
| bug,auto-transferred,priority: normal,category: build/install,affected: 2.4 | low | Critical |
97,435,499 | opencv | calcHist() should raise error for incompatible nimages and dims | Transferred from http://code.opencv.org/issues/3918
```
|| Baris Demiroz on 2014-09-25 14:19
|| Priority: Low
|| Affected: 2.4.9 (latest release)
|| Category: imgproc, video
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: Any / Any
```
## calcHist() should raise error for incompatible nimages and dims
```
When @dims=1@ is passed to @calcHist()@ doesn't take @nimages@ parameter into account, and processes as if @nimages=1@. However, this is not directly or intuitively visible from the documentation. Documentation should be more clear about this behavior and/or @calcHist()@ should raise an error via @CV_Assert()@ for incompatible parameters. The first impression of the function signature is that it is going to aggregate the histograms of all the given images.
PS. I found the whole @calcHist()@ usage obscure, it is very hard to understand and use unlike most of the OpenCV.
```
## History
##### Vladislav Vinogradov on 2014-09-30 12:00
```
Hello Baris,
Could you add a sample code to reproduce the issue?
- Operating System changed from Windows to Any
- HW Platform changed from x64 to Any
- Assignee set to Baris Demiroz
- Status changed from New to Incomplete
- Category set to imgproc, video
```
| bug,auto-transferred,category: imgproc,priority: low,affected: 2.4 | low | Critical |
97,435,598 | opencv | Android VideoCapture output format is not clarified and is missing native YUV NV21 | Transferred from http://code.opencv.org/issues/3925
```
|| Ayberk Özgür on 2014-09-26 14:48
|| Priority: Normal
|| Affected: None
|| Category: highgui-camera
|| Tracker: Feature
|| Difficulty: Easy
|| PR: https://github.com/Itseez/opencv/pull/3280
|| Platform: None / None
```
## Android VideoCapture output format is not clarified and is missing native YUV NV21
```
In cap_android.cpp, the @CvCapture_Android::retrieveFrame( int outputType )@ function takes the raw YUV NV21 (yuv420sp) frame and converts it to one of greyscale, RGB or RGBA to be output to the outside world (see https://github.com/Itseez/opencv/blob/master/modules/videoio/src/cap_android.cpp#L428). This is tied to the @channel@ parameter of @VideoCapture::retrieve(cv::Mat,int)@ which can take one of @CV_CAP_ANDROID_GREY_FRAME@, @CV_CAP_ANDROID_COLOR_FRAME_BGR@, @CV_CAP_ANDROID_COLOR_FRAME_RGB@, @CV_CAP_ANDROID_COLOR_FRAME_BGRA@ or @CV_CAP_ANDROID_COLOR_FRAME_RGBA@. This has two problems:
1) Passing the output type to @VideoCapture@ as @channel@ is very unintuitive. It should at least be passed as a @CV_CAP_@ parameter to the @CvCapture_Android@ object itself through @VideoCapture@, something like @CV_CAP_PROP_ANDROID_OUTPUT_TYPE@. If there's a good reason for passing it as a channel, then it should be properly documented.
2) YUV NV21 type output is missing. There are good reasons for having it as a possible output type, e.g avoiding color space conversions in CPU altogether when only drawing it on screen via a GL shader. YUV NV21 output support could easily be added in ~15 lines of code since it is the default color space of the raw camera image and it is the only space that is guaranteed to be supported by all Android hardware.
```
## History
##### Ilya Lavrenov on 2014-09-27 08:24
```
- Assignee set to Alexander Smorkalov
```
##### Ayberk Özgür on 2014-09-29 13:12
```
PR opened for this: https://github.com/Itseez/opencv/pull/3280
- % Done changed from 0 to 100
```
##### Alexander Smorkalov on 2014-09-30 09:47
```
- Pull request set to https://github.com/Itseez/opencv/pull/3280
```
##### Philip L on 2015-05-24 12:27
```
- % Done changed from 100 to 0
- Status changed from New to Open
```
| auto-transferred,priority: normal,feature,category: videoio(camera) | low | Minor |
97,435,667 | opencv | MOG2_Applay test scenario does not look realistic. | Transferred from http://code.opencv.org/issues/3927
```
|| Konstantin Rodyushkin on 2014-09-29 06:40
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: ocl
|| Tracker: Bug
|| Difficulty: Easy
|| PR:
|| Platform: Any / Any
```
## MOG2_Applay test scenario does not look realistic.
```
MOG2_applay test resets BG/FG statistics every 5 frames that is not realistic. In real scenario FG/BG statistic is collected a long period of time and reset rare in unexpected cases. As result BackgroundSubstructionMOG2.applay performance data gathered by the test is not for real usage case. That makes it unusable.
It is needed to significantly increase processed frames number before statistic is reset.
```
## History
##### Konstantin Rodyushkin on 2014-09-29 07:03
```
Also, the recreation MOG object and statistic reset is rare operations but it is included int measurement scope. So this test is for "Create MOG + Applay" operation but not for Applay
```
##### Ilya Lavrenov on 2014-09-30 14:42
```
- Assignee changed from Vadim Pisarevsky to Alexander Karsakov
- Category changed from imgproc, video to ocl
```
| bug,auto-transferred,priority: normal,affected: 3.4,category: ocl | low | Critical |
97,435,749 | opencv | Highgui_Video.prop_fps test fails on Ubuntu 14.04 with FFMPEG | Transferred from http://code.opencv.org/issues/3930
```
|| Alexander Smorkalov on 2014-09-30 09:37
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: highgui-video
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Linux
```
## Highgui_Video.prop_fps test fails on Ubuntu 14.04 with FFMPEG
```
FFmpeg reports -nan fps value on Ubuntu 14.04 with default version of FFmpeg.
The same test works ok with older FFmpeg from Ubuntu 12.04.
```
## History
| bug,auto-transferred,priority: normal,category: videoio,affected: 3.4 | low | Critical |
97,435,835 | opencv | Normalization of 16UC histogram | Transferred from http://code.opencv.org/issues/3937
```
|| Alex Stax on 2014-10-03 07:40
|| Priority: Normal
|| Affected: None
|| Category: imgproc, video
|| Tracker: Feature
|| Difficulty:
|| PR:
|| Platform: None / None
```
## Normalization of 16UC histogram
```
Currently there is only 8UC histogram normalization functionality.
Could you please add an 16UC implementation?
Thank you
```
## History
##### Steven Puttemans on 2014-10-03 08:15
```
Hello Alex,
Keep in mind that this is an open source library which highly depends on contributions of its user group.
Are you maybe up in adapting the 8 bit code to 16 bit and see how far it gets you?
If you succeed you could supply a PR with the functionality.
This will be the fastest way to get it into the library.
Steven
- Assignee set to Alex Stax
- Status changed from New to Open
```
##### Vadim Pisarevsky on 2015-05-25 11:21
```
- Category set to imgproc, video
```
| auto-transferred,priority: normal,feature,category: imgproc,category: video | low | Minor |
97,435,876 | opencv | HDR Imaging - assert failure during merge_mertens | Transferred from http://code.opencv.org/issues/3939
```
|| Jo Z on 2014-10-05 08:58
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: photo
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x86 / Windows
```
## HDR Imaging - assert failure during merge_mertens
```
According to the HDR-Tutorial: http://docs.opencv.org/trunk/doc/tutorials/photo/hdr_imaging/hdr_imaging.html#hdrimaging
Merge_Mertens does not work.
+Example Code:+
cv::Mat fusion;
cv::Ptr<cv::MergeMertens> merge_mertens = cv::createMergeMertens();
merge_mertens->process(images, fusion);
+Execution results always in the following error:+
OpenCV Error: Assertion failed (y == 0 || (data && dims >= 1 && (unsigned)y < (u nsigned)size.p[0])) in cv::Mat::ptr, file C:\builds\master_PackSlave-win64-vc11- static\opencv\modules\core\include\opencv2/core/mat.inl.hpp, line 750
+The bug happens in this line of code (example: 7 images in the vector, crash at image 4):+
for(int lvl = 0; lvl < maxlevel; lvl++) {
Mat up;
pyrUp(img_pyr[lvl + 1], up, img_pyr[lvl].size());
img_pyr[lvl] -= up;
}
Images are not huge so memory consumption should not be a problem.
```
## History
##### Steven Puttemans on 2014-10-06 08:30
```
Hello!
Before processing that vector of images, could you try doing something dumb and loop over each vector position and imshow the images.
Your error seems due to the fact that there is no data where it expects data to be.
Regards,
Steven
- Assignee changed from Leonid Beynenson to Jo Z
- Status changed from New to Open
```
##### Maksim Shabunin on 2015-04-27 09:11
```
- Target version changed from 3.0-alpha to 3.0
```
##### Vadim Pisarevsky on 2015-05-14 16:42
```
there is regression test on MergeMertens, which passes on different platforms. Without particular sample code and data we could not reproduce the problem and can not fix it.
- Status changed from Open to Incomplete
```
| bug,auto-transferred,priority: normal,affected: 3.4,category: photo | low | Critical |
97,435,910 | opencv | automated ruby bindings for opencv | **Blocked** by #11666
---
Transferred from http://code.opencv.org/issues/3949
```
|| Alexander Duda on 2014-10-11 12:23
|| Priority: Normal
|| Affected: None
|| Category: None
|| Tracker: Feature
|| Difficulty: Easy
|| PR:
|| Platform: None / None
```
## automated ruby bindings for opencv
```
I am wondering if there is an interest in having automated ruby bindings for OpenCV as a OpenCV module.
At the moment I am maintaining ropencv (http://www.ropencv.aduda.eu) which is a ruby gem using the python parser to automatically generate a C layer for all OpenCV C++ methods and a ruby ffi integration script which mimics the C++ interface. The binding is around for over a year and compatible to OpenCV >= 2.4.4. If there is an interest I can integrate it into OpenCV and make a pull request.
```
## History
##### Steven Puttemans on 2014-10-11 13:40
```
Hello! If you are up to making a PR, adding decent documentation and maybe some ruby examples on how to use the bindings, then I am sure people will like the effort!
However, keep in mind that you will have to adapt the bindings to the 3.0 code base, since extra modules are only allowed in master branch and at the opencv_contrib repository.
- Status changed from New to Open
```
##### Alexander Duda on 2014-10-13 13:06
```
What would be the preferred way of adding the module "ruby" (master branch or opencv_contrib)?
```
##### Steven Puttemans on 2014-10-13 14:07
```
Definately the contrib module. No new functionality will be accepted inside the current master branch unless opencv developers decide to integrate it from the contrib repository when a stable and well research modules has been achieved.
```
##### Alexander Duda on 2014-10-18 12:43
```
pull request: https://github.com/Itseez/opencv_contrib/pull/116
```
| auto-transferred,feature,category: build/install,effort: ∞ | low | Minor |
97,436,027 | opencv | VideoWriter_GPU is broken since NVIDIA driver 344.11 | Transferred from http://code.opencv.org/issues/3955
```
|| Sergey Ravdin on 2014-10-14 08:50
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: gpu (cuda)
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Windows
```
## VideoWriter_GPU is broken since NVIDIA driver 344.11
```
New drivers from NVIDIA for Windows completely dropped support for NVCUVENC for NVENC. This means that current implementation of VideoWriter_GPU is useless and produces access violation errors in NVCreateHWEncoder() in video_writer.cpp in VideoWriter_GPU constructor.
Earlier versions of driver lacked nvcuvenc.dll, but VideoWriter_GPU still worked if that dll is placed into system. Now it doesn't.
```
## History
| bug,auto-transferred,priority: normal,affected: 2.4,category: gpu/cuda (contrib) | low | Critical |
97,436,209 | opencv | Issues with shared library naming when using mangled paths | Transferred from http://code.opencv.org/issues/3967
```
|| Gautier Minster on 2014-10-21 10:33
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: build/install
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: Any / Linux
```
## Issues with shared library naming when using mangled paths
```
Hello,
When compiling OpenCV using mangled paths, using versions such as @3.0.0-dev@ or @3.0.0-alpha@, two problems arise:
1. The resulting pkg-config file (eg. @opencv-3.0.0-dev.pc@) refers to the shared libraries using the following format: <pre>libopencv_core.so.3.0.0-dev</pre>
This file does not exist however, the shared libraries are named without the "-dev": <pre>libopencv_core.so.3.0.0</pre>
This of course prevents pkg-config from working.
2. The namelinks of the shared libraries are set: on my system, with a working version of OpenCV 2.4.9, a symlink <pre>libopencv_core.so -> libopencv_core.so.2.4</pre> existed.
Upon installing OpenCV 3.0.0-dev, this symlink was replaced by <pre>libopencv_core.so -> libopencv_core.so.3.0</pre>, breaking OpenCV 2.4.9's pkg-config configuration.
Attached is a patch which:
- fixes the output name of the shared objects when using mangled paths, by adding the OPENCV_VERSION_STATUS to them
- adds the NAMELINK_SKIP flag to the CMake install of shared libraries when using mangled paths, so that already installed namelinks do not get overwritten
Do these 2 changes seem acceptable?
Kind regards,
--
Gautier
```
## History
##### Steven Puttemans on 2014-10-21 12:30
```
Personnally I would not change the use of the namelinks. Now it is fairly easy to build different OpenCV fixes, add a single `pkg-config opencv --libs` to your compiler settings and then just apply a `sudo make install` into the build folder you want to be running. This way you never have to adapt your OpenCV software projects to try out different installations.
```
##### Gautier Minster on 2014-10-23 06:17
```
You would not use mangled paths at all then, and simply install the specific build of OpenCV you want to use at this point? (since when using mangled paths, the pkg-config OpenCV configuration file has a mangled name too)
(note: i deleted the assignee, not sure if that was a mistake on my part or if it was auto-assigned at issue creation)
- Assignee deleted (Roman Donchenko)
```
| bug,auto-transferred,priority: normal,category: build/install,affected: 3.4 | low | Critical |
97,436,254 | opencv | Add eclipse PATH setup details to the Java tutorial | Transferred from http://code.opencv.org/issues/3980
```
|| Phil W on 2014-10-26 12:43
|| Priority: Low
|| Affected: branch '2.4' (2.4-dev)
|| Category: documentation
|| Tracker: Bug
|| Difficulty: Easy
|| PR:
|| Platform: Any / Windows
```
## Add eclipse PATH setup details to the Java tutorial
```
The Eclipse Setup [1] documentation seems to lack the info to setup the PATH variable needed to use certain functionality.
Suggestion:
Add "C:\path\to\opencv\dlls" to the Windows PATH variable or use the Environment tab in Eclipse as shown on the screenshot.
1: http://docs.opencv.org/trunk/doc/tutorials/introduction/java_eclipse/java_eclipse.html
```
## History
| bug,auto-transferred,priority: low,category: documentation,affected: 2.4 | low | Critical |
97,436,296 | opencv | solvePnP: is not handling all degenerated cases | Transferred from http://code.opencv.org/issues/3984
```
|| Alexander Duda on 2014-10-29 10:47
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: calibration, 3d
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: Any / Any
```
## solvePnP: is not handling all degenerated cases
```
Depending on the point configuration and the selected method solvePnP returns wrong solutions
(deviation > 1-e3)
1000 Tests with four fixed scene points (A;B,C,D) and random camera positions
case 1 right angle between ABC; all scene points are coplanar:
SOLVEPNP_ITERATIVE - 0 wrong solutions
SOLVEPNP_P3P - 445 wrong solutions
SOLVEPNP_EPNP - 0 wrong solutions
SOLVEPNP_DLS - 4 wrong solutions
SOLVEPNP_UPNP - 1000 wrong solutions
case 2 right angle between ABC (D is not coplanar to ABC);
SOLVEPNP_ITERATIVE - 634 wrong solutions
SOLVEPNP_P3P - 397 wrong solutions
SOLVEPNP_EPNP - 884 wrong solutions
SOLVEPNP_DLS - 0 wrong solutions
SOLVEPNP_UPNP - 1000 wrong solutions
case 3 (general configuration)
SOLVEPNP_ITERATIVE - 701 wrong solutions
SOLVEPNP_P3P - 0 wrong solutions
SOLVEPNP_EPNP - 613 wrong solutions
SOLVEPNP_DLS - 0 wrong solutions
SOLVEPNP_UPNP - 1000 wrong solutions
case 4 (general configuration + 2 additional points)
SOLVEPNP_ITERATIVE - 0 wrong solutions
SOLVEPNP_P3P -
SOLVEPNP_EPNP - 0 wrong solutions
SOLVEPNP_DLS - 0 wrong solutions
SOLVEPNP_UPNP - 6 wrong solutions
The main problem might be a lack of documentation because P3P is only handling the main case and should only be used in combination with RANSAC and environments having only a few right angles.
```
## History
| bug,auto-transferred,priority: normal,category: calib3d,affected: 2.4 | low | Critical |
97,436,334 | opencv | Android openCV library fails to setup camera on Odroid U3(Samsung GalaxyS3 plateform) or runs extremely slowly | Transferred from http://code.opencv.org/issues/3987
```
|| Alexandre Benoit on 2014-10-30 10:35
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: android
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: ARM / Android
```
## Android openCVlibrary fails to setup camera on Odroid U3(Samsung GalaxyS3 plateform) or runs extremely slowly
```
Hi,
experimenting OpenCV on Android Devices, i encounter troubles on the Odroid U3 plateform (Exynox 4412, GalaxyS3 like, also saw some issues on galaxyS3 in the list). I run my test on the latest OpenCV stable version 2.4.9.
Two points on it that are maybe connected :
_Camera setup fails in JavaCameraView::initializeCamera on line:
@frameSize = calculateCameraFrameSize(sizes, new JavaCameraSizeAccessor(), width, height);@
After this call, frame size contains 0*0 making the following code failing, maybe a value check would protect the following to crash in such case.
_if we force framesize by replacing this line by the following example:
@Size frameSize = new Size(640,480);@
Then, it runs BUT framerate is extremely low : around 1.2fps.
It is currently difficult to locate the origin of the problem : OpenCV or Host Android on the odroid plateform.
I tested with 2 android versions on this plateform : Android 4.4.2 and 4.4.4, with Dalvik or ART, with and without CPU scaling, always getting the same observations.
The same machine can run linux and in this case, everything goes fine and fast with the same peripherals.
These problems has also been observed on odroid XU system (Exynox Octa cpu).
If somebody has an idea of the probem, Odroid team encouraged coming to you but it may come from the OS config (but what?).
Any help will be appreciated.
Thanks !
```
## History
| bug,auto-transferred,priority: normal,affected: 2.4,platform: android | low | Critical |
97,436,539 | opencv | ImageIO reading/writing image on OSX & iOS not supported anymore | Transferred from http://code.opencv.org/issues/4002
```
|| Arnaud Brejeon on 2014-11-08 01:30
|| Priority: Normal
|| Affected: None
|| Category: ios
|| Tracker: Feature
|| Difficulty:
|| PR:
|| Platform: None / None
```
## ImageIO reading/writing image on OSX & iOS not supported anymore
```
I noticed reading/writing image based on OSX & iOS ImageIO is not supported anymore.
Files grfmt_imageio.hpp/cpp have not been moved to the new module
```
## History
##### Vadim Pisarevsky on 2014-11-14 14:09
```
We found that the images loaded through ImageIO are different from the ones loaded through libjpeg/libpng. It causes problems with cross-platform reproduction of the results.
If there is big need in putting this code back, there should be runtime option (set to "libjpeg/libpng" by default).
- Tracker changed from Bug to Feature
```
##### Vadim Pisarevsky on 2015-05-25 11:19
```
- Category set to ios
```
| auto-transferred,priority: normal,feature,platform: ios/osx | low | Critical |
97,436,668 | opencv | Lucas-Kanade Tutorial not working on Python using Mac. | Transferred from http://code.opencv.org/issues/4008
```
|| Quang Nguyen on 2014-11-14 10:12
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: samples
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Mac OSX
```
## Lucas-Kanade Tutorial not working on Python using Mac.
```
http://docs.opencv.org/trunk/doc/py_tutorials/py_video/py_lucas_kanade/py_lucas_kanade.html#lucas-kanade
It breaks down when calling cv2.calcOpticalFlowPyrLK(). lktracker.py in the samples also does not work and outputs the same error.
Output:
OpenCV Error: Image step is wrong (Step must be a multiple of esz1) in setSize, file /tmp/opencv-P6BF4O/modules/core/src/matrix.cpp, line 297
libc++abi.dylib: terminating with uncaught exception of type cv::Exception: /tmp/opencv-P6BF4O/modules/core/src/matrix.cpp:297: error: (-13) Step must be a multiple of esz1 in function setSize
Abort trap: 6
```
## History
##### Quang Nguyen on 2014-11-16 06:36
```
I just installed openCV through anaconda instead of brew and this fixes the error.
First I installed the anaconda package:
http://continuum.io/downloads#all
Then I installed the openCV package via terminal: "conda install opencv"
After that reupdating numpy with: "conda update numpy"
It works now I also see that anaconda is using openCV 2.4.8.
OpenCV 2.4.9 through brew did not work with the samples either, it produced another kind of error in matrix.cpp.
So everything still works in openCV 2.4.8 on Python in Mac but it breaks down in the versions after that.
```
| bug,auto-transferred,priority: normal,category: samples,affected: 3.4 | low | Critical |
97,436,712 | opencv | a bug in the stitching example, caused by unsuitable camera's view | Transferred from http://code.opencv.org/issues/4015
```
|| yajn lv on 2014-11-19 10:37
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: stitching
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x86 / Windows
```
## a bug in the stitching example, caused by unsuitable camera's view
```
Sometimes fail to allocate memory. It is caused by the bundle adjustment function provide unsuitable camera's rotation matrix. When the input images are just in the front of eyes, rotation matrices should be like [1,0,0;0,1,0;0,0,1], but [1,0,0;0,0,1;0,-1,0] was provided.
I multiply the matrices by one of it's inverse to prevent this problem,but it only works on my present work. Is there any method to dual with it?
```
## History
##### Vadim Pisarevsky on 2015-04-27 15:20
```
- Category set to stitching
```
| bug,auto-transferred,priority: normal,affected: 2.4,category: stitching | low | Critical |
97,436,764 | opencv | CvVideoCamera flickering iOS | Transferred from http://code.opencv.org/issues/4017
```
|| Ashwin Prasad on 2014-11-20 18:26
|| Priority: High
|| Affected: branch 'master' (3.0-dev)
|| Category: ios
|| Tracker: Bug
|| Difficulty: Medium
|| PR:
|| Platform: ARM / iOS
```
## CvVideoCamera flickering iOS
```
The CvVideoCamera is causing the camera capture to flicker on iOS devices using OpenCV 3.0 beta.
The same code was working fine with OpenCV 2.4.9
//ViewController.h
#import "opencv2/videoio/cap_ios.h"
@interface ViewController : UIImageView<CvVideoCameraDelegate>
@property (nonatomic, strong) CvVideoCamera* videoCamera;
//ViewController.m
self.videoCamera = [[CvVideoCamera alloc] initWithParentView:self];
self.videoCamera.defaultAVCaptureDevicePosition = AVCaptureDevicePositionFront;
self.videoCamera.defaultAVCaptureSessionPreset = AVCaptureSessionPreset640x480;
self.videoCamera.defaultAVCaptureVideoOrientation = AVCaptureVideoOrientationPortrait;
self.videoCamera.defaultFPS = 30;
self.videoCamera.delegate = self;
self.videoCamera.parentView = self;
```
## History
##### Maksim Shabunin on 2015-04-27 09:12
```
- Target version changed from 3.0-beta to 3.0
```
| bug,auto-transferred,affected: 3.4,platform: ios/osx | low | Critical |
97,436,844 | opencv | No windowedMatchingMask() function in opencv3 | Transferred from http://code.opencv.org/issues/4022
```
|| Ozan Caglayan on 2014-11-24 14:27
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: features2d
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: Any / Any
```
## No windowedMatchingMask() function in opencv3
```
Hi,
I was trying to port an existing OpenCV code to the 3.0-beta but figured out that the function windowedMatchingMask() which was defined in modules/features2d/src/matchers.cpp in OpenCV 2.x no longer exists in OpenCV 3.
I think it was actually moved into the contrib repository's xfeatures2d module with the following commit:
commit d4a77fc42858fc3711e156115c76318d93b7ccb0
Author: Vadim Pisarevsky <[email protected]>
Date: Mon Aug 11 23:25:30 2014 +0400
added xfeatures2d (made of opencv/nonfree and a part of opencv/features2d)
But apparently the above commit didn't bring that function into opencv_contrib and commented out the single call to windowedMatchingMask():
modules/xfeatures2d/samples/video_homography.cpp: //Mat mask = windowedMatchingMask(test_kpts, train_kpts, 25, 25);
So maybe the function was dropped completely but I couldn't find any clue about this on Google.
```
## History
##### Maksim Shabunin on 2015-04-27 09:12
```
- Target version changed from 3.0-beta to 3.0
```
| bug,auto-transferred,priority: normal,category: features2d,affected: 3.4 | low | Critical |
97,437,005 | opencv | BackgroundSubtractorMOG2 apply function fails ASSERT for large frame sizes | Transferred from http://code.opencv.org/issues/4027
```
|| Matt K on 2014-11-25 15:47
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: t-api
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Linux
```
## BackgroundSubtractorMOG2 apply function fails ASSERT for large frame sizes
```
Reproduce this by using the
opencv3/samples/cpp/tutorial_code/video/bg_sub.cpp
with a video of 4450 x 2000 frame size.
The line
pMOG2->apply(frame, fgMaskMOG2);
returns
OpenCV Error: Assertion failed (retval == CL_SUCCESS) in _allocateBufferEntry, file /home/opencv/opencv3/modules/core/src/ocl.cpp, line 3533
OpenCV Error: Assertion failed (u != 0) in create, file /home/opencv/opencv3/modules/core/src/umatrix.cpp, line 275
terminate called after throwing an instance of 'cv::Exception'
what(): /home/opencv/opencv3/modules/core/src/umatrix.cpp:275: error: (-215) u != 0 in function create
OpenCV version 3.0.0b
The sample works ok if one resizes the frame to smaller such as 1400x625 before calling pMOG2->apply(frame, fgMaskMOG2);
System is on 3.13.0-29-generic #53-Ubuntu SMP Wed Jun 4 21:00:20 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux
MemTotal: 16359600 kB
SwapTotal: 39062524 kB
VmallocTotal: 34359738367 kB
HugePages_Total: 0
Video card AMD Radeon R9 200 Series 2048MB with latest OpenCL runtime
Not sure if this is a memory issue although memory requirement for pMOG2->apply for a frame size is not reported and I did not try estimate it.
Background removal from 4k camera may be considered an interesting application for opencv
Thanks!
Matt
```
## History
##### Ilya Lavrenov on 2014-12-01 11:39
```
Hi Matt,
I agree with this bug. If you don't need the OpenCL version, you can set `setUseOpenCL(false)` and MOG2 will be CPU-only (because currently it copies the data to GPU and use OpenCL version which is faster)
```
##### Vadim Pisarevsky on 2015-04-27 15:13
```
- Category set to t-api
```
| bug,auto-transferred,priority: normal,affected: 3.4,category: t-api | low | Critical |
97,437,046 | opencv | Java bindings. MatOfPoint default constructor creates Mat with CV_8UC1 instead of CV_32S2 type. | Transferred from http://code.opencv.org/issues/4029
```
|| Vladimir Protsenko on 2014-11-26 13:13
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: java bindings
|| Tracker: Bug
|| Difficulty: Easy
|| PR:
|| Platform: x64 / Any
```
## Java bindings. MatOfPoint default constructor creates Mat with CV_8UC1 instead of CV_32S2 type.
```
```
## History
##### Ilya Lavrenov on 2014-12-01 11:32
```
- Target version set to 2.4.11
- Difficulty set to Easy
- Priority changed from Low to Normal
- Assignee set to Andrey Pavlenko
- Status changed from New to Open
- Category set to java bindings
```
| bug,auto-transferred,priority: normal,affected: 2.4,category: java bindings | low | Critical |
97,437,080 | opencv | Compile fails with unicap on debian wheezy 32 bits | Transferred from http://code.opencv.org/issues/4032
```
|| Vasco Santos on 2014-11-27 00:14
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: build/install
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x86 / Linux
```
## Compile fails with unicap on debian wheezy 32 bits
```
-- OpenCV modules:
-- To be built: core flann imgproc highgui features2d calib3d ml video legacy objdetect photo gpu ocl nonfree contrib java python stitching superres ts videostab
-- Disabled: world
-- Disabled by dependency: -
-- Unavailable: androidcamera dynamicuda viz
--
-- GUI:
-- QT: NO
-- GTK+ 2.x: YES (ver 2.24.10)
-- GThread : YES (ver 2.32.4)
-- GtkGlExt: NO
-- OpenGL support: NO
-- VTK support: NO
--
-- Media I/O:
-- ZLib: build (ver 1.2.7)
-- JPEG: /usr/lib/i386-linux-gnu/libjpeg.so (ver 80)
-- PNG: /usr/lib/i386-linux-gnu/libpng.so (ver 1.2.49)
-- TIFF: /usr/lib/i386-linux-gnu/libtiff.so (ver 42 - 3.9.6)
-- JPEG 2000: /usr/lib/i386-linux-gnu/libjasper.so (ver 1.900.1)
-- OpenEXR: /usr/lib/libImath.so /usr/lib/libIlmImf.so /usr/lib/libIex.so /usr/lib/libHalf.so /usr/lib/libIlmThread.so (ver 1.6.1)
--
-- Video I/O:
-- DC1394 1.x: NO
-- DC1394 2.x: YES (ver 2.2.0)
-- FFMPEG: YES
-- codec: YES (ver 53.35.0)
-- format: YES (ver 53.21.1)
-- util: YES (ver 51.22.2)
-- swscale: YES (ver 2.1.0)
-- gentoo-style: YES
-- GStreamer:
-- base: YES (ver 0.10.36)
-- app: YES (ver 0.10.36)
-- video: YES (ver 0.10.36)
-- OpenNI: NO
-- OpenNI PrimeSensor Modules: NO
-- PvAPI: NO
-- GigEVisionSDK: NO
-- UniCap: YES (ver 0.9.12)
-- UniCap ucil: YES (ver 0.9.5)
-- V4L/V4L2: Using libv4l (ver 0.8.8)
-- XIMEA: NO
-- Xine: NO
--
-- Other third-party libraries:
-- Use IPP: NO
-- Use Eigen: NO
-- Use TBB: NO
-- Use OpenMP: NO
-- Use GCD NO
-- Use Concurrency NO
-- Use C=: NO
-- Use Cuda: NO
-- Use OpenCL: YES
--
-- OpenCL:
-- Version: dynamic
-- Include path: /home/user/Projects/sandbox/opencv-2.4.9/3rdparty/include/opencl/1.2
-- Use AMD FFT: NO
-- Use AMD BLAS: NO
--
-- Python:
-- Interpreter: /usr/bin/python2 (ver 2.7.3)
-- Libraries: /usr/lib/libpython2.7.so (ver 2.7.3)
-- numpy: /usr/lib/pymodules/python2.7/numpy/core/include (ver 1.6.2)
-- packages path: lib/python2.7/dist-packages
--
-- Java:
-- ant: /usr/bin/ant (ver 1.8.2)
-- JNI: /usr/lib/jvm/default-java/include /usr/lib/jvm/default-java/include /usr/lib/jvm/default-java/include
-- Java tests: YES
--
-- Documentation:
-- Build Documentation: NO
-- Sphinx: NO
-- PdfLaTeX compiler: /usr/bin/pdflatex
--
-- Tests and samples:
-- Tests: YES
-- Performance tests: YES
-- C/C++ Examples: YES
--
-- Install path: /usr/local
--
-- cvconfig.h is in: /home/user/Projects/sandbox/opencv-2.4.9/release
-- -----------------------------------------------------------------
--
-- Configuring done
-- Generating done
CMake Warning:
Manually-specified variables were not used by the project:
BUILD_PYTHON_SUPPORT
-- Build files have been written to: /home/user/Projects/sandbox/opencv-2.4.9/release
(..) shortened (..)
Scanning dependencies of target opencv_highgui
[ 9%] Building CXX object modules/highgui/CMakeFiles/opencv_highgui.dir/src/cap.cpp.o
[ 9%] Building CXX object modules/highgui/CMakeFiles/opencv_highgui.dir/src/cap_images.cpp.o
[ 9%] Building CXX object modules/highgui/CMakeFiles/opencv_highgui.dir/src/cap_ffmpeg.cpp.o
[ 9%] Building CXX object modules/highgui/CMakeFiles/opencv_highgui.dir/src/loadsave.cpp.o
[ 9%] Building CXX object modules/highgui/CMakeFiles/opencv_highgui.dir/src/utils.cpp.o
[ 9%] Building CXX object modules/highgui/CMakeFiles/opencv_highgui.dir/src/window.cpp.o
[ 10%] Building CXX object modules/highgui/CMakeFiles/opencv_highgui.dir/src/window_gtk.cpp.o
[ 10%] Building CXX object modules/highgui/CMakeFiles/opencv_highgui.dir/src/cap_dc1394_v2.cpp.o
[ 10%] Building CXX object modules/highgui/CMakeFiles/opencv_highgui.dir/src/cap_gstreamer.cpp.o
[ 10%] Building CXX object modules/highgui/CMakeFiles/opencv_highgui.dir/src/cap_unicap.cpp.o
In file included from /home/user/Projects/sandbox/opencv-2.4.9/modules/highgui/src/cap_unicap.cpp:46:0:
/usr/include/unicap/unicap.h:218:34: error: expected unqualified-id before ‘private’
/usr/include/unicap/unicap.h:218:33: error: expected ‘;’ at end of member declaration
/usr/include/unicap/unicap.h:218:41: error: expected ‘:’ before ‘;’ token
make[2]: *** [modules/highgui/CMakeFiles/opencv_highgui.dir/src/cap_unicap.cpp.o] Error 1
make[1]: *** [modules/highgui/CMakeFiles/opencv_highgui.dir/all] Error 2
make: *** [all] Error 2
```
## History
##### Vadim Pisarevsky on 2015-04-27 15:12
```
- Category set to build/install
```
| bug,auto-transferred,priority: normal,category: build/install,affected: 2.4 | low | Critical |
97,437,106 | opencv | cv::stereoRectify returns incorrect Q matrix | Transferred from http://code.opencv.org/issues/4035
```
|| Gerald Lodron on 2014-12-01 12:06
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: calibration, 3d
|| Tracker: Bug
|| Difficulty: Easy
|| PR:
|| Platform: x64 / Windows
```
## cv::stereoRectify returns incorrect Q matrix
```
cv::stereoRectify computes the projection matrix Q which is used in reprojectImageTo3D to get the real 3d points from disparities. To get correct ones we need to invert Q(3,3):
Q(3,3)=-Q(3,3)
before calling reprojectImageTo3D
I tested it in opencv 2.4.9
I searched a while for a reason of my wrong 3d point cloud and found the solution on a 4 years old bug report in ROS:
https://code.ros.org/trac/opencv/ticket/898
would be nice if this bug could get fixed since it is very old....
```
## History
##### Vadim Pisarevsky on 2015-04-27 15:12
```
- Priority changed from High to Normal
- Category set to calibration, 3d
```
| bug,auto-transferred,category: calib3d,affected: 2.4,incomplete | low | Critical |
97,437,135 | opencv | OpenCV Cheat Sheet in Python | Transferred from http://code.opencv.org/issues/4036
```
|| Eduard Feicho on 2014-12-01 12:50
|| Priority: Normal
|| Affected: None
|| Category: documentation
|| Tracker: Feature
|| Difficulty:
|| PR:
|| Platform: None / None
```
## OpenCV Cheat Sheet in Python
```
Hi,
I think it would be really useful to have something like the OpenCV Cheat Sheet, but instead of C++ having it in Python.
I find myself quite often confused about how to do things using numpy, so it would be a great help for the basic stuff.
```
## History
##### Steven Puttemans on 2014-12-01 13:26
```
+1 for this! Would definately be an improvement. What about auto generating cheatsheets for `Java / Python / Matlab`? Is that possible? Afaik the cheatsheet is manually created now and then by a volunteer.
```
| auto-transferred,priority: normal,feature,category: documentation | low | Major |
97,437,163 | opencv | OpenCV Error in blender->prepare(corners, sizes) | Transferred from http://code.opencv.org/issues/4039
```
|| Thorsten Behrens on 2014-12-02 10:52
|| Priority: Low
|| Affected: branch 'master' (3.0-dev)
|| Category: t-api
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Windows
```
## OpenCV Error in blender->prepare(corners, sizes)
```
Hi all,
I receive the folloing error, when the ROI created with blender->prepare(corners, sizes) as in the stitching_detailed.cpp is too big:
_OpenCV Error: Assertion failed (clEnqueueReadBuffer(q, (cl_mem)u->handle, CL_TRUE, 0, u->size, alignedPtr.getAlignedPtr(), 0, 0, 0) == CL_SUCCESS) in map, file C:\opencv3.0\sources\modules\core\src\ocl.cpp, line 3961
terminate called after throwing an instance of 'cv::Exception'
what(): C:\opencv3.0\sources\modules\core\src\ocl.cpp:3961: error: (-215) clEnqueueReadBuffer(q, (cl_mem)u->handle, CL_TRUE, 0, u->size,alignedPtr.getAlignedPtr(), 0, 0, 0) == CL_SUCCESS in function map_
The problen occurs somewhere between 6500*6500 pixel and 10500*10500 pixel.
Any fix is greatly appreciated!
TIA and best regards,
Thorsten
```
## History
##### Vadim Pisarevsky on 2015-04-27 15:10
```
without sample code (do you use the standard opencv sample?) _and_ the input dataset we can not reproduce and thus can not solve the problem
- Priority changed from Normal to Low
```
##### Vadim Pisarevsky on 2015-04-27 15:10
```
- Category set to t-api
```
| bug,auto-transferred,priority: low,affected: 3.4,category: t-api | low | Critical |
97,437,501 | opencv | Crash in flann (kdtree_index.h method searchLevelExact) | Transferred from http://code.opencv.org/issues/4073
```
|| Sebastian Bartmann on 2014-12-18 09:15
|| Priority: High
|| Affected: branch 'master' (3.0-dev)
|| Category: flann
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Windows
```
## Crash in flann (kdtree_index.h method searchLevelExact)
```
Under certain circumstances the third parameter(const NodePtr node) passed in the searchLevelExact method in file kdtree_index.h seems to be NULL, leading to a crash.
The method traverses the tree recursively. Unfortunately it isn't checked that a child is really existing before trying to visit it. The method only stops traversing if both childs aren't present.
I'm not sure if a null pointer check is missing here or if it is a tree construction problem. Maybe a node should always have two children or none. In this case the method would be right.
```
## History
| bug,auto-transferred,affected: 3.4,category: flann | low | Critical |
97,437,536 | opencv | Create a module to read/write AVIs using Directshow instead of VFW | Transferred from http://code.opencv.org/issues/4075
```
|| Christophe Caltagirone on 2014-12-18 21:59
|| Priority: Normal
|| Affected: None
|| Category: highgui-video
|| Tracker: Feature
|| Difficulty:
|| PR:
|| Platform: None / None
```
## Create a module to read/write AVIs using Directshow instead of VFW
```
The current OpenCV implementation for reading and writing AVIs is based on Video for Windows, which does not support files over 2GB. It would be nice to create a capture plugin to read/write AVIs based on DirectShow, which supports OpenDML AVI formats allowing bigger files.
```
## History
| auto-transferred,priority: normal,feature,category: videoio | low | Minor |
97,437,571 | opencv | the fuction cvCreateFileCapture return 0 on Opencv 2.4.9 | Transferred from http://code.opencv.org/issues/4078
```
|| 若然 陈 on 2014-12-19 10:07
|| Priority: Low
|| Affected: 2.4.9 (latest release)
|| Category: highgui-video
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Windows
```
## the fuction cvCreateFileCapture return 0 on Opencv 2.4.9
```
The function failed and cannot play video when I was opening the AVI File, So I Stepped into the function by VS2010 debugger,and found out that it just called the cvCreateFileCapture_FFMPEG_proxy funtion return 0(failed). then it called cvCreateFileCapture_Images this function,What a pity is that cvCreateFileCapture_Images also failed.Especially,if the filename contains chinese character ,the icvExtractPattern would crash in Debug mode.
But, The function succeed and play video sucessfully if I was using Opencv 2.4.6 version.
Is it a Bug of opencv?
```
## History
##### Artur Wieczorek on 2014-12-30 20:31
```
From what you've described it seems that your library is built with HAVE_VFW and HAVE_MSMF options disabled.
Please try to enable these options when building the library and check if this helps.
If not, please provide a link to the sample avi file causing the problem (or attach the smallest possible part of it).
```
##### Vadim Pisarevsky on 2015-04-27 13:25
```
- Priority changed from Normal to Low
- Category set to highgui-video
```
| bug,auto-transferred,priority: low,category: videoio,affected: 2.4 | low | Critical |
97,437,754 | opencv | pyopen wont get imported unless numpy is imported | Transferred from http://code.opencv.org/issues/4089
```
|| Hakim Benoudjit on 2014-12-28 14:36
|| Priority: Low
|| Affected: 2.4.9 (latest release)
|| Category: python bindings
|| Tracker: Bug
|| Difficulty: Easy
|| PR:
|| Platform: x64 / Other
```
## pyopen wont get imported unless numpy is imported
```
Hi,
My issue is that when I enter in the `python` console:
_>>> import cv2_
I get the following error:
_ImportError: numpy.core.multiarray failed to import
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
ImportError: numpy.core.multiarray failed to import_
unless I import numpy before importing py-opencv:
_>>> import numpy
>>> import cv2_
I'm using Freebsd 10 and here are the versions of the libs used:
Python: 2.7.8
Numpy: 1.9.1
Py-Opencv: 2.4.9
```
## History
##### Vadim Pisarevsky on 2015-04-27 13:23
```
probably, it's system dependent behaviour. On OSX and Linux "import cv2" works just fine. Therefore, it's unlikely we will fix the bug ourselves.
- Category set to python bindings
```
| bug,auto-transferred,category: python bindings,priority: low,affected: 2.4 | low | Critical |
97,437,815 | opencv | Canny can not detect connected contour of black square on a white background. | Transferred from http://code.opencv.org/issues/4093
```
|| Vit Shiryaev on 2014-12-31 14:26
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: ocl
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x86 / Windows
```
## Canny can not detect connected contour of black square on a white background.
```
Result of Simple code for simple image is not connected contour as it is expected, top or bottom edge missed depending on aperture parameter. Result printscreen is !http://answers.opencv.org/upfiles/14199661605456809.png! (see source image attached).
Code is:
<pre>
UMat img = imread("./square.png", IMREAD_COLOR).getUMat(ACCESS_READ);
UMat imgGray, imgGrayEdges3, imgGrayEdges5;
cvtColor(img, imgGray, CV_BGR2GRAY);
Canny(imgGray, imgGrayEdges3, 50, 150, 3);
Canny(imgGray, imgGrayEdges5, 300, 2000, 5); // why so large thresholds?..
imshow("img", img);
imshow("canny3", imgGrayEdges3);
imshow("canny5", imgGrayEdges5);
waitKey();
</pre>
_I use OpenCV3-beta in Eclipse under Win8.1 and have to build source from git. Source updated and compiled today (last canny.cpp change was 7 weeks ago). I do have openCL so I guess UMat Canny function uses OCL_RUN version of function. According to some reports ( http://answers.opencv.org/question/52114/?answer=52128#post-id-52128 ) there are no such behaviour in 2.4_
```
## History
##### Nisarg Thakkar on 2015-01-08 06:31
```
Works for me. I get all the four sides. I'm using the latest version available on github.
```
##### Vit Shiryaev on 2015-01-08 11:23
```
Nisarg Thakkar, updated just now from github master, error remains the same. What is your cmake config (see my attached)?
- File cmake-gui-out.txt added
```
##### Claudio Caraffi on 2015-01-08 16:15
```
Any possibility that it is a visualization aliasing problem? What about if you dump the image and open it with another visualization tool?
```
##### Vit Shiryaev on 2015-01-08 17:07
```
Claudio Caraffi, OpenCV function imshow(...) always use 100% scale thus no reason to do aliasing. I checked it before by "cout << imgGray.getMat(ACCESS_READ) << "\n" << imgGrayEdges3.getMat(ACCESS_READ)" as you can see from link to answers.opencv.org in report above. But I checked it with imwrite(..) and another visualization tool -- result is predictably the same.
```
##### Vit Shiryaev on 2015-01-12 21:18
```
I have tested this code for 2.4 on the same PC (using oclMat, see http://answers.opencv.org/question/52114/ comments), Canny works fine, no visible bugs.
```
##### Claudio Caraffi on 2015-01-13 12:38
```
I could reproduce the same bug on my machine. Equivalent OpenCL code works fine with OpenCV 2.4, has some inconsistencies in OpenCV 3.0. I tried with different "real" images, and compared the cv::Mat and cv::UMat versions, depending on the kernel size I have:
kernel 3: about 0.5% of edges missing (like 1500 pixels out of 300K).
kernel 5: about 0.1% of edges missing
kernel 7: something new appears, and something is missing, for a total of about 0.1% pixel different
I would change the title of this bug to "OpenCL Canny inconsistent with standard Canny"
OpenCV 3 built with MinGW4.9.2 32bit, IPP disabled, TBB enabled.
Note: Canny in version 3.0 is 4 times faster of version 2.9 (both OpenCL)!
```
##### Claudio Caraffi on 2015-01-14 14:09
```
I have created related issue http://code.opencv.org/issues/4120
```
##### Maksim Shabunin on 2015-04-27 09:12
```
- Target version changed from 3.0-beta to 3.0
```
##### Maksim Shabunin on 2015-04-27 10:58
```
- Category set to ocl
```
| bug,auto-transferred,priority: normal,affected: 3.4,category: ocl | low | Critical |
97,437,910 | opencv | Adding support for Metal language under iOS platform. | Transferred from http://code.opencv.org/issues/4100
```
|| Ricardo Ruiz on 2015-01-07 15:44
|| Priority: Low
|| Affected: None
|| Category: ios
|| Tracker: Feature
|| Difficulty: Hard
|| PR:
|| Platform: None / None
```
## Adding support for Metal language under iOS platform.
```
Hello!
As you know, we don't have OpenCL under iOS unless you jailbreak your device.
It seems Apple don't like OpenCL. However they created its own language called Metal and it seems they are very similar:
i.imgur.com/K858yLi.jpg
Here docs about Metal:
https://developer.apple.com/library/ios/documentation/Metal/Reference/MetalFrameworkReference/
My request would be support this language because probably artificial vision will improve a lot in this mobile platform.
```
## History
| auto-transferred,feature,priority: low,platform: ios/osx | low | Major |
97,438,033 | opencv | imwrite can't write jpeg2000 4-channel images. | Transferred from http://code.opencv.org/issues/4111
```
|| Nisarg Thakkar on 2015-01-08 18:21
|| Priority: Low
|| Affected: branch '2.4' (2.4-dev)
|| Category: highgui-images
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Linux
```
## imwrite can't write jpeg2000 4-channel images.
```
Load a png with an alpha channel. Use CV_LOAD_IMAGE_UNCHANGED. Try to write it to a jp2 image. imwrite fails returning 0.
```
## History
##### Nisarg Thakkar on 2015-02-22 04:19
```
Currently jasper doesn't support 4 channel images.
```
##### Vadim Pisarevsky on 2015-04-28 11:27
```
- Priority changed from Normal to Low
- Category changed from core to highgui-images
```
| bug,auto-transferred,priority: low,affected: 2.4,category: imgcodecs | low | Critical |
97,438,150 | opencv | Cannot get gpu opencv 2.4.10 to work. Cuda 6.5, Intel C++ 15. GCC 4.6-4.9, Ubuntu 14.10 | Transferred from http://code.opencv.org/issues/4115
```
|| Steph van Schalkwyk on 2015-01-12 04:47
|| Priority: Normal
|| Affected: branch '2.4' (2.4-dev)
|| Category: gpu (cuda)
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Linux
```
## Cannot get gpu opencv 2.4.10 to work. Cuda 6.5, Intel C++ 15. GCC 4.6-4.9, Ubuntu 14.10
```
GTX 760, Ubuntu 14.10, tried Intel c++ 15 (with CUDA hack), GCC from 4.6 to 4.9.
"cmake -D CMAKE_BUILD_TYPE=RELEASE -D CMAKE_INSTALL_PREFIX=/usr/local -D WITH_CUBLAS=ON -D WITH_CUFFT=ON -D WITH_EIGEN=ON -D WITH_OPENGL=ON -D WITH_QT=ON -D WITH_TBB=ON -D BUILD_DOCS=ON -D BUILD_EXAMPLES=ON -D BUILD_TESTS=ON -D CUDA_ARCH_BIN="3.0" .."
Purged system of 2.4.9 but seems "/build/buildd/opencv-2.4.9+dfsg..." (below) still there. Also openGlut problems on openCV examples.
"OpenCV Error: No GPU support (The library is compiled without CUDA support) in mallocPitch, file /build/buildd/opencv-2.4
.9+dfsg/modules/dynamicuda/include/opencv2/dynamicuda/dynamicuda.hpp, line 126
terminate called after throwing an instance of 'cv::Exception'
what(): /build/buildd/opencv-2.4.9+dfsg/modules/dynamicuda/include/opencv2/dynamicuda/dynamicuda.hpp:126: error: (-216
) The library is compiled without CUDA support in function mallocPitch"
```
## History
##### Ilya Lavrenov on 2015-01-13 22:33
```
- Assignee set to Vladislav Vinogradov
```
##### Ilya Lavrenov on 2015-01-13 22:34
```
Hi Steph,
Please, provide a CMake output.
```
##### Vadim Pisarevsky on 2015-04-27 13:13
```
- Priority changed from Blocker to Normal
- Category set to gpu (cuda)
```
##### Philip L on 2015-04-28 18:56
```
I think you missed the "WITH_CUDA" flag
-D WITH_CUDA=ON
should be added
```
##### Steven Puttemans on 2015-04-29 08:04
```
Philip L wrote:
> I think you missed the "WITH_CUDA" flag
> -D WITH_CUDA=ON
> should be added
If you add the CUFFT and CUBLAS flag ON, then the CUDA option is enabled by default, since they depend on building CUDA :)
```
##### Philip L on 2015-04-30 15:44
```
ok since i am using the gui for setting up things like this i thought this would be the issue. Ok this could also be true due to the way the gui will handle the cmake call.
```
| bug,auto-transferred,priority: normal,affected: 2.4,category: gpu/cuda (contrib) | low | Critical |
97,438,297 | opencv | OpenCL SURF_OCL detecting no features with Intel Processor Graphics | Transferred from http://code.opencv.org/issues/4121
```
|| Michael Gopshtein on 2015-01-15 12:50
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: ocl
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: Any / Windows
```
## OpenCL SURF_OCL detecting no features with Intel Processor Graphics
```
I'm using version 2.4.10. When I perform "ocl::getOpenCLDevices" I get the following devices listed:
0: "Intel(R) Core(TM) i7-4810MQ CPU @ 2.80GHz:Intel(R) Corporation"
1: "Intel(R) HD Graphics 4600:Intel(R) Corporation"
2: "Quadro K2100M:NVIDIA Corporation"
I tried selecting all three devices with "ocl::setDevice()". Next I'm running "ocl::SURF_OCL" algorithm to detect keypoint of given image. When I select devices 0 (main CPU) or 2 (Nvidia card) it seems to work fine although providing different results.
Once I select device #1 (the Intel GPU card embedded on the CPU), same code on same picture detects 0 keypoints.
Project was compiled with all of "WITH_OPENCL", "WITH_OPENCLLAMDBLAS" and "WITH_OPENCLAMDFFT" selected, Visual Studio 2014. Both 32 and 64 bit.
Thanks,
Michael
```
## History
##### Michael Gopshtein on 2015-01-15 12:53
```
It's Visual Studio 2013 (not 2014)
```
##### Vadim Pisarevsky on 2015-04-27 13:08
```
- Category set to ocl
```
| bug,auto-transferred,priority: normal,affected: 2.4,category: ocl | low | Critical |
97,438,374 | opencv | ocl::BFMatcher_OCL::knnMatch fails when trained first [2.4.10] | Transferred from http://code.opencv.org/issues/4132
```
|| Michael Gopshtein on 2015-01-18 08:33
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: ocl
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x86 / Windows
```
## ocl::BFMatcher_OCL::knnMatch fails when trained first [2.4.10]
```
In the following block the call to "knnMatch" fails:
<pre>
ocl::BFMatcher_OCL matcher;
vector< vector<DMatch> > matches;
std::vector<ocl::oclMat> trainm(1);
trainm[0] = descriptors1;
matcher.add(trainm);
matcher.knnMatch(descriptors2, matches, 2);
</pre>
Without "train" it works:
<pre>
matcher.knnMatch(descriptors2, descriptors1, matches, 2);
</pre>
Exception:
* _interlockedExchangeAdd [Line 899]
* cv::ocl::oclMat::release() [Line 601]
* cv::ocl::oclMat::operator=() [Line 244]
* cv::ocl::BruteForceMatcher_OCL_base::makeGpuCollection() [Line 635]
* cv::ocl::BruteForceMatcher_OCL_base::knnMatch() [Line 944]
Works fine on 64bit. Visual Studio 2013.
Full code:
<pre>
int main(int argc, char *argv[]) {
ocl::DevicesInfo devs;
ocl::getOpenCLDevices(devs, ocl::CVCL_DEVICE_TYPE_ALL);
ocl::setDevice(devs[2]); // Quadro K2100M : NVIDIA Corporation
Mat img1, img1BW, img2, img2BW;
img1 = imread("C:\\Work\\Projects\\OpenCVFeatureMatching\\images\\pizza1.png");
img2 = imread("C:\\Work\\Projects\\OpenCVFeatureMatching\\images\\pizza2.png");
cvtColor(img1, img1BW, COLOR_RGB2GRAY);
cvtColor(img2, img2BW, COLOR_RGB2GRAY);
ocl::oclMat oclImg1(img1BW);
ocl::oclMat oclImg2(img2BW);
ocl::oclMat points1, points2; // THIS LINE WORKS
ocl::SURF_OCL detector1;
ocl::SURF_OCL detector2;
ocl::oclMat descriptors1, descriptors2;
detector1(oclImg1, ocl::oclMat(), points1, descriptors1);
detector2(oclImg2, ocl::oclMat(), points2, descriptors2);
ocl::BFMatcher_OCL matcher;
vector< vector<DMatch> > matches;
std::vector<ocl::oclMat> trainm(1);
trainm[0] = descriptors1;
matcher.add(trainm);
matcher.knnMatch(descriptors2, matches, 2);
return 0;
}
</pre>
Thanks
```
## History
| bug,auto-transferred,priority: normal,affected: 2.4,category: ocl | low | Critical |
97,438,493 | opencv | opecv_contrib xfeatures2d cuda surf compile error | Transferred from http://code.opencv.org/issues/4140
```
|| Samuel Petrocelli on 2015-01-21 16:32
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: gpu (cuda)
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Linux
```
## opecv_contrib xfeatures2d cuda surf compile error
```
[ 93%] Building CXX object modules/xfeatures2d/CMakeFiles/opencv_xfeatures2d.dir/src/surf.cuda.cpp.o
/usr/local/src/opencv_contrib/modules/xfeatures2d/src/surf.cuda.cpp: In constructor ‘{anonymous}::SURF_CUDA_Invoker::SURF_CUDA_Invoker(cv::cuda::SURF_CUDA&, const cv::cuda::GpuMat&, const cv::cuda::GpuMat&)’:
/usr/local/src/opencv_contrib/modules/xfeatures2d/src/surf.cuda.cpp:149:59: error: invalid initialization of reference of type ‘cv::cuda::Stream&’ from expression of type ‘cv::cuda::GpuMat’
cuda::integral(img, surf_.sum, surf_.intBuffer);
^
In file included from /usr/local/src/opencv/build/modules/xfeatures2d/precomp.hpp:63:0:
/usr/local/src/opencv/modules/cudaarithm/include/opencv2/cudaarithm.hpp:699:17: error: in passing argument 3 of ‘void cv::cuda::integral(cv::InputArray, cv::OutputArray, cv::cuda::Stream&)’
CV_EXPORTS void integral(InputArray src, OutputArray sum, Stream& stream = Stream::Null());
^
/usr/local/src/opencv_contrib/modules/xfeatures2d/src/surf.cuda.cpp:155:75: error: invalid initialization of reference of type ‘cv::cuda::Stream&’ from expression of type ‘cv::cuda::GpuMat’
cuda::integral(surf_.mask1, surf_.maskSum, surf_.intBuffer);
^
In file included from /usr/local/src/opencv/build/modules/xfeatures2d/precomp.hpp:63:0:
/usr/local/src/opencv/modules/cudaarithm/include/opencv2/cudaarithm.hpp:699:17: error: in passing argument 3 of ‘void cv::cuda::integral(cv::InputArray, cv::OutputArray, cv::cuda::Stream&)’
CV_EXPORTS void integral(InputArray src, OutputArray sum, Stream& stream = Stream::Null());
^
make[2]: *** [modules/xfeatures2d/CMakeFiles/opencv_xfeatures2d.dir/src/surf.cuda.cpp.o] Error 1
make[1]: *** [modules/xfeatures2d/CMakeFiles/opencv_xfeatures2d.dir/all] Error 2
make: *** [all] Error 2
[root@spetroce build]# gedit /usr/local/src/opencv_contrib/modules/xfeatures2d/src/surf.cuda.cpp
^C
[root@spetroce build]# gedit /usr/local/src/opencv_contrib/modules/xfeatures2d/src/surf.cuda.cpp
(gedit:25630): Gtk-WARNING **: Calling Inhibit failed: GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name org.gnome.SessionManager was not provided by any .service files
(gedit:25630): Gtk-WARNING **: Calling Inhibit failed: GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name org.gnome.SessionManager was not provided by any .service files
```
## History
##### Tomasz Pietruszka on 2015-01-29 21:36
```
Just wanted to confirm it happening,
in my case x86 architecture, ubuntu, eg.:
<pre>
cmake -D CMAKE_BUILD_TYPE=RELEASE -DOPENCV_EXTRA_MODULES_PATH=../../opencv_contrib/modules -DBUILD_EXAMPLES=ON -DBUILD_PERF_TESTS=OFF -DWITH_IPP=OFF -DCUDA_ARCH_BIN="2.0" ..
</pre>
```
##### Nicolas Thorstensen on 2015-02-03 15:47
```
I know how to fix this. Where should I post the solution?
```
##### Steven Puttemans on 2015-02-04 07:52
```
Nicolas Thorstensen wrote:
> I know how to fix this. Where should I post the solution?
Of course you should. And better yet, supply a PR with the fix!
```
##### Steven Puttemans on 2015-02-04 07:52
```
- Priority changed from Blocker to Normal
- Status changed from New to Open
```
##### Vadim Pisarevsky on 2015-04-27 12:58
```
- Category set to gpu (cuda)
```
| bug,auto-transferred,priority: normal,affected: 3.4,category: gpu/cuda (contrib) | low | Critical |
97,438,522 | opencv | Extension of fastNlMeansDenoising to accept 16-bit inputs | Transferred from http://code.opencv.org/issues/4143
```
|| Iliya Romm on 2015-01-22 11:27
|| Priority: Normal
|| Affected: None
|| Category: photo
|| Tracker: Feature
|| Difficulty:
|| PR:
|| Platform: None / None
```
## Extension of fastNlMeansDenoising to accept 16-bit inputs
```
Certain cameras used in research applications (in my case - an infrared camera) output grayscale images of up to 16-bit (in my case, 14). Currently the application of the filter on these data requires quantization to 8 bits, which results in the loss of vital experimental data. It would therefore be very beneficial to have 16-bit support to allow application of the filter on the original "more than 8"-bit data.
*****
I can contribute to the task by testing the output against equivalent MATLAB CPU implementations (my knowledge in C\C++ is simply not sufficient to do this on my own, yet my considerably better MATLAB knowledge can be put to use).
```
## History
##### Iliya Romm on 2015-01-22 11:38
```
Can somebody please change the category of this to "photo"?
```
##### Erik Karlsson on 2015-02-13 13:29
```
I have a basic 16-bit implementation up and running. It can be found here.
https://github.com/erikrk/opencv/tree/denoising-16bit
Currently, I have the following issues:
1) A lookup table (almost_dist2weight_) is used to calculate e^(-d2/h²). This works fine for 8 bits, but with 16 bits you will run out of memory if h is too big.
2) fastNlMeansDenoisingColored and fastNlMeansDenoisingColoredMulti do not work since cvtColor does not support 16 bits for COLOR_LBGR2Lab and COLOR_Lab2LBGR.
```
##### Erik Karlsson on 2015-02-18 21:23
```
I have added two new functions, fastNlMeansDenoisingAbs and fastNlMeansDenoisingMultiAbs, which use squared sum of absolute value distances instead of sum of squared distances to calculate weights. The code can be found here, in the branch denoising-16bit-abs.
https://github.com/erikrk/opencv/tree/denoising-16bit-abs
They support both 8- and 16-bit 1-, 2-, and 3-channel data.
The original fuctions (fastNlMeansDenoising and fastNlMeansDenoisingMulti) are still 8-bit only. The code has been extensivley refactored though in order to facilitete addition of 16-bit support. It does pass all regression tests. Extending those functions to 16 bits should be straightforward as long as you can come up with a fast metod for calculating weight = e^(-d2/(nchannels*h^2)) that doses not use lookup tables which become too big for squared 16-bit distances.
For colored denoising to accept 16-bit input, cvtColor would have to be extended to accept 16-bit input for COLOR_LBGR2Lab and COLOR_Lab2LBGR.
- % Done changed from 0 to 80
- Status changed from New to Incomplete
```
##### Vadim Pisarevsky on 2015-05-25 12:45
```
- Category set to photo
```
| auto-transferred,feature,priority: low,category: photo | low | Minor |
97,438,555 | opencv | Warning Target "opencv_videoio" requests linking to directory | Transferred from http://code.opencv.org/issues/4147
```
|| Albert Nava on 2015-01-24 00:41
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: highgui-video
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Windows
```
## Warning Target "opencv_videoio" requests linking to directory
```
I have a trouble in Cmake after it's finished the process of Configure, I start the process of Generate, but I got the following errors :
WARNING: Target "opencv_videoio" requests linking to directory "C:/Program Files (x86)/OpenNI/Lib". Targets may link only to libraries. CMake is dropping the item.
I'm using windows 8.1 with python 2.7
So I don`t know what is the problem. Or what does Cmake search and where..
Could you please give me some idea of what is wrong .
Thanks and regards
```
## History
##### Vadim Pisarevsky on 2015-04-27 12:56
```
try to build opencv without OpenNI support, there seems to be some problem with cmake script or with the options that you passed to it
- Priority changed from High to Normal
- Category set to highgui-video
```
| bug,auto-transferred,priority: normal,category: videoio,affected: 3.4 | low | Critical |
97,438,618 | opencv | CMake inter-target dependency graph contains a cycle | Transferred from http://code.opencv.org/issues/4157
```
|| Gregory Kramida on 2015-01-29 15:06
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: build/install
|| Tracker: Bug
|| Difficulty: Easy
|| PR:
|| Platform: x64 / Linux
```
## CMake inter-target dependency graph contains a cycle
```
When trying to generate, I'm getting the following CMake error with CMake 3.1.1 on Ubuntu:
@CMake Error: The inter-target dependency graph contains the following strongly connected component (cycle):
"opencv_python3" of type SHARED_LIBRARY
depends on "opencv_python2" (weak)
"opencv_python2" of type SHARED_LIBRARY
depends on "opencv_python3" (weak)
At least one of these targets is not a STATIC_LIBRARY. Cyclic dependencies are allowed only among static libraries.@
```
## History
##### Gregory Kramida on 2015-01-29 15:09
```
Work-around: turn off either BUILD_opencv_python2 or BUILD_opencv_python3
```
##### Ilya Lavrenov on 2015-02-01 18:28
```
- Assignee set to Maksim Shabunin
```
##### Nico Yep on 2015-02-09 17:16
```
Hello, I can confirm this bug on Archlinux.
```
##### Maksim Shabunin on 2015-03-26 07:33
```
Gregory, can you please check if this issue is reproduced with latest master branch revision? Also, please, provide cmake flags that you used.
```
##### Maksim Shabunin on 2015-04-27 09:12
```
- Target version changed from 3.0-beta to 3.0
```
##### Maksim Shabunin on 2015-04-27 09:33
```
Can not reproduce this issue with Ubuntu 14.04, cmake 3.1.1, python 2.7 + 3.4, latest master branch.
Please, provide more information about your system configuration and cmake command line parameters.
- Status changed from New to Incomplete
- Category set to build/install
```
| bug,auto-transferred,priority: normal,category: build/install,affected: 3.4 | low | Critical |
97,438,677 | opencv | segment_objects.cpp cannot work. | Transferred from http://code.opencv.org/issues/4160
```
|| Sheng Liu on 2015-01-30 03:34
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: imgproc, video
|| Tracker: Bug
|| Difficulty: Hard
|| PR:
|| Platform: x64 / Windows
```
## segment_objects.cpp cannot work.
```
The file segment_objects.cpp ，which provided in the opencv3.0beta，cannot run .Could you help me solve the problem?
What's more,I need write a program.The program can realize that tracking a moving or static object in a vedio.The object's size is small.And the vedio camera is not always static.It' a very difficult work for me.Could you give me some advice?Is there some similar program I can learn?Please give me some links to the program.
```
## History
| bug,auto-transferred,priority: normal,category: imgproc,category: video,affected: 3.4 | low | Critical |
97,438,744 | opencv | Bad Display Aspect Ratio for videos | Transferred from http://code.opencv.org/issues/4163
```
|| Pierre Letessier on 2015-01-30 14:18
|| Priority: Normal
|| Affected: None
|| Category: highgui-video
|| Tracker: Feature
|| Difficulty: Easy
|| PR:
|| Platform: None / None
```
## Bad Display Aspect Ratio for videos
```
I have videos with a SAR (aspect ratio) != 1, see below the ffprobe info :
<pre>
Stream #0:0[0x8ad]: Video: mpeg2video (Main) ([2][0][0][0] / 0x0002), yuv420p, 720x576 [SAR 64:45 DAR 16:9], 25 fps, 25 tbr, 90k tbn, 50 tbc
</pre>
FFMpeg or vlc can both play them with the good display aspect ratio (16/9), but not OpenCV, so when I load them, the videos are stretched, and I have to resize the grabbed frames.
Furthermore, there are not any functions available to know the DAR.
Could it be possible to add a "CV_CAP_PROP_DAR" property and the corresponding code in opencv-2.4.6.1/modules/highgui/src/cap_ffmpeg_impl.hpp ?
<pre>
case CV_FFMPEG_CAP_PROP_DAR:
AVRational sar = ic->streams[video_stream]->sample_aspect_ratio;
double dar = (double)frame.width * (double) sar.num / ((double)frame.height * (double) sar.den);
return dar;
</pre>
I did a patch in OpenCV 2.4.6.1 (see attachment)
Or directly read the video with the good aspect ratio ?
Thank you !
```
## History
| auto-transferred,priority: normal,feature,category: videoio | low | Minor |
97,438,777 | opencv | scope issue in createLaplacePyrGpu() in Blenders.cpp | Transferred from http://code.opencv.org/issues/4164
```
|| Barry Secrest on 2015-02-01 14:03
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: gpu (cuda)
|| Tracker: Bug
|| Difficulty: Easy
|| PR:
|| Platform: Any / Any
```
## scope issue in createLaplacePyrGpu() in Blenders.cpp
```
Original code:
pyr.resize(num_levels + 1);
vector<gpu::GpuMat> gpu_pyr(num_levels + 1);
gpu_pyr[0].upload(img);
for (int i = 0; i < num_levels; ++i)
gpu::pyrDown(gpu_pyr[i], gpu_pyr[i + 1]);
gpu::GpuMat tmp;
for (int i = 0; i < num_levels; ++i)
{
gpu::pyrUp(gpu_pyr[i + 1], tmp);
gpu::subtract(gpu_pyr[i], tmp, gpu_pyr[i]);
}
gpu_pyr[i].download(pyr[i]);
This then causes a compile error since "i" is no longer in scope.
Corrected code:
int i;
pyr.resize(num_levels + 1);
vector<gpu::GpuMat> gpu_pyr(num_levels + 1);
gpu_pyr[0].upload(img);
for (i = 0; i < num_levels; ++i)
gpu::pyrDown(gpu_pyr[i], gpu_pyr[i + 1]);
gpu::GpuMat tmp;
for (i = 0; i < num_levels; ++i)
{
gpu::pyrUp(gpu_pyr[i + 1], tmp);
gpu::subtract(gpu_pyr[i], tmp, gpu_pyr[i]);
}
gpu_pyr[i].download(pyr[i]);
```
## History
##### akansh agrawal on 2015-02-02 08:53
```
The exact code for the above code in Blenders.cpp is:
pyr.resize(num_levels + 1);
std::vector<cuda::GpuMat> gpu_pyr(num_levels + 1);
gpu_pyr[0].upload(img);
for (int i = 0; i < num_levels; ++i)
cuda::pyrDown(gpu_pyr[i], gpu_pyr[i + 1]);
cuda::GpuMat tmp;
for (int i = 0; i < num_levels; ++i)
{
cuda::pyrUp(gpu_pyr[i + 1], tmp);
cuda::subtract(gpu_pyr[i], tmp, gpu_pyr[i]);
gpu_pyr[i].download(pyr[i]);
}
gpu_pyr[num_levels].download(pyr[num_levels]);
- Assignee set to abid rahman
```
##### akansh agrawal on 2015-02-02 08:55
```
akansh agrawal wrote:
> The exact code for the above code in Blenders.cpp is:
>
> pyr.resize(num_levels + 1);
>
> std::vector<cuda::GpuMat> gpu_pyr(num_levels + 1);
> gpu_pyr[0].upload(img);
> for (int i = 0; i < num_levels; ++i)
> cuda::pyrDown(gpu_pyr[i], gpu_pyr[i + 1]);
>
> cuda::GpuMat tmp;
> for (int i = 0; i < num_levels; ++i)
> {
> cuda::pyrUp(gpu_pyr[i + 1], tmp);
> cuda::subtract(gpu_pyr[i], tmp, gpu_pyr[i]);
> gpu_pyr[i].download(pyr[i]);
> }
>
> gpu_pyr[num_levels].download(pyr[num_levels]);
- Assignee changed from abid rahman to Barry Secrest
```
##### Vadim Pisarevsky on 2015-04-27 12:53
```
- Category set to gpu (cuda)
```
| bug,auto-transferred,priority: normal,affected: 2.4,category: gpu/cuda (contrib) | low | Critical |
97,438,805 | opencv | createLaplacePyrGpu() doesnt work | Transferred from http://code.opencv.org/issues/4165
```
|| Barry Secrest on 2015-02-01 14:52
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: gpu (cuda)
|| Tracker: Bug
|| Difficulty: Medium
|| PR:
|| Platform: x64 / Windows
```
## createLaplacePyrGpu() doesnt work
```
From blenders.cpp:
if (can_use_gpu_ && img_with_border.depth() == CV_16S)
createLaplacePyrGpu(img_with_border, num_bands_, src_pyr_laplace);
else createLaplacePyr(img_with_border, num_bands_, src_pyr_laplace);
My attempt at using the simple stitching example was failing (I have CUDA) with the src_pyr_laplace[] returning without initialized data pointers.
I changed blenders.cpp to:
/* if (can_use_gpu_ && img_with_border.depth() == CV_16S)
createLaplacePyrGpu(img_with_border, num_bands_, src_pyr_laplace);
else */
createLaplacePyr(img_with_border, num_bands_, src_pyr_laplace);
and everything worked.
```
## History
##### Vadim Pisarevsky on 2015-04-27 12:53
```
- Category set to gpu (cuda)
```
| bug,auto-transferred,priority: normal,affected: 2.4,category: gpu/cuda (contrib) | low | Critical |
97,438,873 | opencv | cv::cuda::multiply and cv::cuda::sqr can't handle large vector matrices. | Transferred from http://code.opencv.org/issues/4168
```
|| Daniel Moodie on 2015-02-03 13:52
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: gpu (cuda)
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Windows
```
## cv::cuda::multiply and cv::cuda::sqr can't handle large vector matrices.
```
MSVC 2013, Win7x64, OpenCV 3.0 (built from source against cuda 6.5)
I have a large 560,000+ element long row matrix. I'm trying to square it with either cv::cuda::sqr or cv::cuda::multiply. When I try to do this I get:
include\\opencv2\\cudev\\grid\\detail/transform.hpp:270: error: (-217) invalid configuration argument in functi...
The exact same code works fine until the matrix grows to be larger than about 560,000 elements.
```
## History
| bug,auto-transferred,priority: normal,affected: 3.4,category: gpu/cuda (contrib) | low | Critical |
97,439,087 | opencv | Camera Status Check | Transferred from http://code.opencv.org/issues/4188
```
|| Muhammad Qaisar on 2015-02-10 08:59
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: highgui-video
|| Tracker: Bug
|| Difficulty: Medium
|| PR:
|| Platform: x86 / Windows
```
## Camera Status Check
```
Dear sir
I am Using opencv 2.4.10
i am capturing video from axis video server. i use this code
int main()
{
cv::VideoCapture cap;
cap.open("rtsp://192.168.0.90:554/mpeg4/media.amp?type=.mjpg");
if(!cap.isOpened())
{
cout <<"cam not working"<<endl;
}
while(cap.isOpened())
{
Mat Frame;
cap.read(Frame);
if(Frame.empty())
{
cout <<""<<endl;
}
waitKey(1);
}
return 0;
}
the code working good, but here the problem is when i unplug the power cable of camera NO VIDEO shows on top left corner. but i did not get any error...
my question is that how can i check the camera status on axis video server if running camera is disconnected... plz give me example with code. thanks
```
## History
##### Vadim Pisarevsky on 2015-04-27 12:45
```
- Category set to highgui-video
```
| bug,auto-transferred,priority: normal,category: videoio,affected: 2.4 | low | Critical |
97,439,117 | opencv | Viz3d crashes when it is closed | Transferred from http://code.opencv.org/issues/4191
```
|| Romain Brégier on 2015-02-12 09:56
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: viz
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Windows
```
## Viz3d crashes when it is closed
```
Hi,
When testing Viz tutorials, the viewer works as expected, but when I quit the viewer, the exit event is triggered at least twice, causing an access violation exception (interactor == NULL for the second call) :
Here is the code called twice : "vizimpl.cpp", line 86 :
=========================================================
void cv::viz::Viz3d::VizImpl::ExitCallback::Execute(vtkObject*, unsigned long event_id, void*)
{
if (event_id == vtkCommand::ExitEvent)
{
//THIS PART OF CODE IS EXECUTED AT LEAST TWICE
viz->interactor_->TerminateApp();
viz->interactor_ = 0;
}
}
I tested the following example code of the tutorial :
http://docs.opencv.org/trunk/doc/tutorials/viz/widget_pose/widget_pose.html#widget-pose
My settings are :
==========
* Windows 7 64bits
* MSVC12
* VTK 6.1 dynamic libraries.
* OpenCV 3.0 beta
commit ae4cb571ab4fdca3c7be57f0de19743f823daf3f
Author: Alexander Alekhin <[email protected]>
Date: Fri Nov 7 19:22:46 2014 +0300
Set status to beta
```
## History
##### Anatoly Baksheev on 2015-02-12 10:43
```
Thanks for reporting. Could you provide a code where you invoke Viz?
```
##### Romain Brégier on 2015-02-12 12:18
```
I performed no modifications in the code snippet shown in the following tutorial :
http://docs.opencv.org/trunk/doc/tutorials/viz/widget_pose/widget_pose.html#widget-pose
(see attached file)
Thanks for your support,
Romain
- File viz.cpp added
```
##### Romain Brégier on 2015-02-12 12:47
```
Here is an even simpler snippet where I :
* create a Viz3D
* create a WCube (I do not call showWidget)
* Call spin()
* assign a variable just in order to put a break point and ensure that up to here, there are no runtime problem
* End
The program crashes at its end, with an "access violation exception" (see screenshot -in French), but caused by a different function than in my previous post :
The destructor of the Widget class
In widget.cpp
================
cv::viz::Widget::~Widget()
{
if (impl_)
{
delete impl_;
// Here the debugger throw me an exception, and I can see impl_->prop == 0
impl_ = 0;
}
}
- File viz_basic_3.cpp added
- File viz-basic_bug.PNG added
```
##### Anatoly Baksheev on 2015-04-21 12:27
```
The widget problem is pretty odd, I think it's a debugger issue and problem is not in Widget class.
I can't this anywas with VTK 5.8/VS2008. It is really strange that the callback is called after spin(), it is expected only after spinOnce(). But might be VTK 6.1 has some behavioral changes, or you use interleaved spin()/spinOnce(). That is not recommended.
Also this PR https://github.com/Itseez/opencv/pull/3947 seems to has workaround of this. Thanks to @tcavallari.
```
##### Maksim Shabunin on 2015-04-27 09:12
```
- Target version changed from 3.0-beta to 3.0
```
| bug,auto-transferred,priority: normal,affected: 3.4,category: viz | low | Critical |
97,439,187 | opencv | FileNode and FileNodeIterator does not work for std::vector<std::string> | Transferred from http://code.opencv.org/issues/4196
```
|| Yusuke Kameda on 2015-02-15 13:41
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: core
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: Any / Any
```
## FileNode and FileNodeIterator does not work for std::vectorstd::string
```
http://docs.opencv.org/modules/core/doc/xml_yaml_persistence.html#id2
Compile error occurs in this code.
<pre>
std::vector<std::string> v;
cv::FileStorage fs;
cv::FileNode fn = fs["a"];
fn >> v; // error
cv::FileNodeIterator fni = fn.begin();
fni >> v; // error
</pre>
```
## History
##### Yusuke Kameda on 2015-02-17 08:43
```
It seems like the code properly works for 'std::vector<cv::String>',
so I think it is not a major issue.
Please cancel this issue.
<pre>
std::vector<cv::String> v;
cv::FileStorage fs;
cv::FileNode fn = fs["a"];
fn >> v;
cv::FileNodeIterator fni = fn.begin();
fni >> v;
</pre>
```
##### Maksim Shabunin on 2015-04-27 09:12
```
- Target version changed from 3.0-beta to 3.0
```
| bug,auto-transferred,priority: normal,category: core,affected: 3.4 | low | Critical |
97,439,227 | opencv | Disconnected IP camera (opencv 2.4.9 and 3.0) | Transferred from http://code.opencv.org/issues/4199
```
|| Martin Šimák on 2015-02-16 18:23
|| Priority: High
|| Affected: 2.4.9 (latest release)
|| Category: highgui-camera
|| Tracker: Bug
|| Difficulty: Easy
|| PR:
|| Platform: x86 / Windows
```
## Disconnected IP camera (opencv 2.4.9 and 3.0)
```
I try to use my camera with opencv 2.4.9 and 3.0
When i use .open() on connected camera it´s fine.
But when i disconnect the camera and then use .open or .read or (VideoCapture)cap>>img or anything else that is working with camera output then program free/stuck/endless loop.
There is no chance to unfreeze without connection of camera. (i want to use multithreading and HttpClient to solve but...this is big BUG)
There is no chance to try the STATUS of camera (connected / disconnected) before using of .open() or .read()
```
## History
##### Steven Puttemans on 2015-02-17 07:35
```
Could you please run a debug session and look which cap interface from the highgui module is used after passing through the cap.cpp source file?
This will help us to identify which grabbing interface is broken and should be fixed.
- Status changed from New to Open
```
##### Martin Šimák on 2015-02-17 09:04
```
Steven Puttemans wrote:
> Could you please run a debug session and look which cap interface from the highgui module is used after passing through the cap.cpp source file?
> This will help us to identify which grabbing interface is broken and should be fixed.
I start Local Windows Debugger but i am not sure if i am doing it right.
When i step into is says i can view only dissasembly code (source information is missing for debug information for this module ) ...in CodeMap i saw it stucks in loop _ZwWaitForSingleObject&12 but i am not sure what i am doing.
If you want more accurate answer i need some help to debug right.
Maybe simmilar problem was http://code.opencv.org/issues/2786
I try to use that cap_ffmpeg_impl but it wont help its old issue
```
##### Martin Šimák on 2015-02-18 20:27
```
Steven Puttemans wrote:
> Could you please run a debug session and look which cap interface from the highgui module is used after passing through the cap.cpp source file?
> This will help us to identify which grabbing interface is broken and should be fixed.
Maybe it is simmilar with WinHTTP.
When i try to use this code with BAD WEBSITE WinHttpSendRequest step into infinite loop...maybe same in opencv.
But in WINHTTP you can do client.SetTimeouts(0U,1000U,1000U); after 1second..it stop
Please help me how to solve bad URL requestin in OPENCV
DWORD dwSize = 0;
DWORD dwDownloaded = 0;
LPSTR pszOutBuffer;
BOOL bResults = FALSE;
HINTERNET hSession = NULL,
hConnect = NULL,
hRequest = NULL;
// Use WinHttpOpen to obtain a session handle.
hSession = WinHttpOpen(L"WinHTTP Example/1.0",
WINHTTP_ACCESS_TYPE_DEFAULT_PROXY,
WINHTTP_NO_PROXY_NAME,
WINHTTP_NO_PROXY_BYPASS, 0);
cout << "1 " << hSession << endl;
WinHttpSetOption(hSession, WINHTTP_OPTION_REDIRECT_POLICY_ALWAYS, NULL, NULL);
// Specify an HTTP server.
if (hSession)
hConnect = WinHttpConnect(hSession, L"www.microsot.com",
INTERNET_DEFAULT_HTTPS_PORT, 0);
cout << "1.2 " << hConnect << endl;
// Create an HTTP request handle.
if (hConnect)
hRequest = WinHttpOpenRequest(hConnect, L"GET", NULL,
NULL, WINHTTP_NO_REFERER,
WINHTTP_DEFAULT_ACCEPT_TYPES,
WINHTTP_FLAG_SECURE);
cout << "1.3 " << hRequest << endl;
// Send a request.
if (hRequest)
bResults = WinHttpSendRequest(hRequest,
WINHTTP_NO_ADDITIONAL_HEADERS, 0,
WINHTTP_NO_REQUEST_DATA, 0,
0, 0);
- Assignee set to Steven Puttemans
```
| bug,auto-transferred,affected: 2.4,category: videoio(camera) | low | Critical |
97,439,259 | opencv | cv.imread() causes Matlab segfault | Transferred from http://code.opencv.org/issues/4201
```
|| Henry Braun on 2015-02-17 18:19
|| Priority: Normal
|| Affected: branch 'master' (3.0-dev)
|| Category: matlab bindings
|| Tracker: Bug
|| Difficulty:
|| PR:
|| Platform: x64 / Mac OSX
```
## cv.imread() causes Matlab segfault
```
I'm running Matlab r2014b on Mac OS 10.10.1. Official OpenCV bindings from opencv_contrib appeared to build and install perfectly.
I am running this line of code:
<pre>
im = cv.imread('peppers.png');
</pre>
This produces a segmentation violation and crashes Matlab. This also happens with other Matlab test images ('cameraman.tif', 'football.jpg') but does not occur when I try it with one of my own images. Copying the test images from within matlab_r2014b.app to my desktop and loading from there does not fix the problem.
Matlab output is below:
<pre>
------------------------------------------------------------------------
Segmentation violation detected at Tue Feb 17 10:53:27 2015
------------------------------------------------------------------------
Configuration:
Crash Decoding : Disabled
Current Visual : Quartz
Default Encoding : ISO-8859-1
MATLAB Architecture: maci64
MATLAB Root : /Applications/MATLAB_R2014b.app
MATLAB Version : 8.4.0.150421 (R2014b)
Operating System : Darwin 14.0.0 Darwin Kernel Version 14.0.0: Fri Sep 19 00:26:44 PDT 2014; root:xnu-2782.1.97~2/RELEASE_X86_64 x86_64
Processor ID : x86 Family 6 Model 69 Stepping 1, GenuineIntel
Software OpenGL : 0
Virtual Machine : Java 1.7.0_55-b13 with Oracle Corporation Java HotSpot(TM) 64-Bit Server VM mixed mode
Window System : Quartz
Fault Count: 1
Abnormal termination:
Segmentation violation
Register State (from fault):
RAX = 0000000000001002 RBX = 0000000000001400
RCX = 000000010f968af0 RDX = 0000000000001400
RSP = 00007fff9c01be68 RBP = 0000000000001002
RSI = 0000000101d3c000 RDI = 0000000000001002
R8 = ff80000000001002 R9 = 0000000000000000
R10 = 000000010f968adc R11 = 000000010f968ae0
R12 = 000000010f968af0 R13 = 0000000134a644f9
R14 = ffffffffffffffff R15 = 00007fff9c01be68
RIP = 00007fff7e4aa0c8 RFL = 0000000000000000
CS = 0000000135095940 FS = 000000010f968f80 GS = 00007fff7ce5d1d8
Stack Trace (from fault):
[ 0] 0x000000010281ae64 /Applications/MATLAB_R2014b.app/bin/maci64/libmwfl.dylib+00028260 _ZN2fl4diag15stacktrace_base7captureERKNS0_14thread_contextEm+00000052
[ 1] 0x000000010281ea2a /Applications/MATLAB_R2014b.app/bin/maci64/libmwfl.dylib+00043562 _ZN2fl4test17terminate_handledEv+00000906
[ 2] 0x000000010281e477 /Applications/MATLAB_R2014b.app/bin/maci64/libmwfl.dylib+00042103 _ZN2fl4diag13terminate_logEPKcPK17__darwin_ucontext+00000119
[ 3] 0x0000000102272b7a /Applications/MATLAB_R2014b.app/bin/maci64/libmwmcr.dylib+00387962 _Z32mnRunPathDependentInitializationv+00003146
[ 4] 0x0000000102272e58 /Applications/MATLAB_R2014b.app/bin/maci64/libmwmcr.dylib+00388696 _Z32mnRunPathDependentInitializationv+00003880
[ 5] 0x0000000102270733 /Applications/MATLAB_R2014b.app/bin/maci64/libmwmcr.dylib+00378675 mnFatalSignalHandler+00000275
[ 6] 0x00007fff8dce5f1a /usr/lib/system/libsystem_platform.dylib+00020250 _sigtramp+00000026
[ 7] 0x5800000102d3b517 <unknown-module>+00000000
[ 8] 0x0000000134a5ba64 /Users/henry/Downloads/opencv-3.0.0-beta/staticLibs/lib/libopencv_core.3.0.dylib+01313380 _ZN2cv9transposeERKNS_11_InputArrayERKNS_12_OutputArrayE+00000052
[ 9] 0x000000012263f409 /usr/local/matlab/+cv/imread.mexmaci64+00238601 _ZN2cv6bridge20deepCopyAndTransposeIhhEEvRKNS_3MatERN6matlab7MxArrayE+00000665
[ 10] 0x0000000122607f9c /usr/local/matlab/+cv/imread.mexmaci64+00012188 _ZN2cv6bridge6Bridge7FromMatIhEEN6matlab7MxArrayERKNS_3MatE+00000204
[ 11] 0x0000000122606e8c /usr/local/matlab/+cv/imread.mexmaci64+00007820 _ZN2cv6bridge6Bridge7FromMatIN6matlab11InheritTypeEEENS3_7MxArrayERKNS_3MatE+00000156
[ 12] 0x00000001226075d1 /usr/local/matlab/+cv/imread.mexmaci64+00009681 mexFunction+00001201
[ 13] 0x00000001072faeeb /Applications/MATLAB_R2014b.app/bin/maci64/libmex.dylib+00069355 mexRunMexFile+00000091
[ 14] 0x00000001072f6f41 /Applications/MATLAB_R2014b.app/bin/maci64/libmex.dylib+00053057 _ZN7Mfh_mex30runMexFileWithSignalProtectionEiPP11mxArray_tagiS2_+00000113
[ 15] 0x00000001072f7130 /Applications/MATLAB_R2014b.app/bin/maci64/libmex.dylib+00053552 _ZN7Mfh_mex13dispatch_fileEiPP11mxArray_tagiS2_+00000304
[ 16] 0x0000000107386c28 /Applications/MATLAB_R2014b.app/bin/maci64/libmwm_dispatcher.dylib+00400424 _ZN8Mfh_file11dispatch_fhEiPP11mxArray_tagiS2_+00000600
[ 17] 0x000000010756ae8e /Applications/MATLAB_R2014b.app/bin/maci64/libmwmcos.dylib+01093262 _ZNK4mcos17MScopedMethodInfo6invokeEPKNS_9COSClientERiPNS_8COSValueEiPKS5_+00000606
[ 18] 0x0000000107508597 /Applications/MATLAB_R2014b.app/bin/maci64/libmwmcos.dylib+00689559 _ZN24OpaqueMethodCallMetaData11callMMethodEPN4mcos13COSMethodInfoEPP11mxArray_tagPKNS0_9COSClientE+00000359
[ 19] 0x000000010750be63 /Applications/MATLAB_R2014b.app/bin/maci64/libmwmcos.dylib+00704099 _ZN24OpaqueMethodCallMetaData23callMethodOrConstructorEP11mxArray_tagbiPS1_iPPKS0_bPKN4mcos9COSClientEPS9_PPNS6_13COSMethodInfoE+00000259
[ 20] 0x000000010750ad41 /Applications/MATLAB_R2014b.app/bin/maci64/libmwmcos.dylib+00699713 _Z22omCallMethodWithClientP11mxArray_tagbS0_iPS0_iPPKS_PKN4mcos9COSClientEPS8_PPNS5_13COSMethodInfoE+00000417
[ 21] 0x0000000107596ee8 /Applications/MATLAB_R2014b.app/bin/maci64/libmwmcos.dylib+01273576 _Z18omCallStaticMethodP11mxArray_tagiPS0_iPPKS_+00000088
[ 22] 0x000000010761fe64 /Applications/MATLAB_R2014b.app/bin/maci64/libmwmcos.dylib+01834596 _ZN17Mfh_opaque_static11dispatch_mfEiPP11mxArray_tagiS2_+00000116
[ 23] 0x000000010732a12d /Applications/MATLAB_R2014b.app/bin/maci64/libmwm_dispatcher.dylib+00020781 _ZN13Mfh_MATLAB_fn11dispatch_fhEiPP11mxArray_tagiS2_+00000333
[ 24] 0x000000010761cacc /Applications/MATLAB_R2014b.app/bin/maci64/libmwmcos.dylib+01821388 _ZN10Mfh_opaque11dispatch_fhEiPP11mxArray_tagiS2_+00000028
[ 25] 0x0000000107ceed62 /Applications/MATLAB_R2014b.app/bin/maci64/libmwm_interpreter.dylib+01523042 _Z19inDispatchFromStackiPKcii+00001650
[ 26] 0x0000000107c90dd8 /Applications/MATLAB_R2014b.app/bin/maci64/libmwm_interpreter.dylib+01138136 inPcodeWalker+00022552
[ 27] 0x0000000107c8d917 /Applications/MATLAB_R2014b.app/bin/maci64/libmwm_interpreter.dylib+01124631 inPcodeWalker+00009047
[ 28] 0x0000000107c88fb7 /Applications/MATLAB_R2014b.app/bin/maci64/libmwm_interpreter.dylib+01105847 inInterPcode+00000935
[ 29] 0x0000000107c8a1c6 /Applications/MATLAB_R2014b.app/bin/maci64/libmwm_interpreter.dylib+01110470 _Z26inExecuteMFunctionOrScriptP6Mfh_mpb+00000886
[ 30] 0x0000000107d2c1f1 /Applications/MATLAB_R2014b.app/bin/maci64/libmwm_interpreter.dylib+01774065 _Z10inRunMfileiPP11mxArray_tagiS1_P6Mfh_mpP15inWorkSpace_tag+00002337
[ 31] 0x0000000107386c55 /Applications/MATLAB_R2014b.app/bin/maci64/libmwm_dispatcher.dylib+00400469 _ZN8Mfh_file11dispatch_fhEiPP11mxArray_tagiS2_+00000645
[ 32] 0x0000000107ccf48b /Applications/MATLAB_R2014b.app/bin/maci64/libmwm_interpreter.dylib+01393803 _Z23inEvalPcodeHeaderToWordP15_memory_contextiPP11mxArray_tagP12_pcodeheaderP6Mfh_mpj+00000219
[ 33] 0x0000000107c81f4b /Applications/MATLAB_R2014b.app/bin/maci64/libmwm_interpreter.dylib+01077067 _Z25in_local_call_with_setjmpIN5boost3_bi6bind_tIvPFvP15_memory_contextPiPP11mxArray_tagP12_pcodeheaderjENS1_5list5INS1_5valueIS4_EENS0_3argILi1EEENSG_ILi2EEENSE_ISA_EENSE_IiEEEEEEEN6mlutil14cmddistributor17inExecutionStatusET_S5_S8_b+00000171
[ 34] 0x0000000107c7d221 /Applications/MATLAB_R2014b.app/bin/maci64/libmwm_interpreter.dylib+01057313 _Z12inEvalStringP15_memory_contextRKSbItSt11char_traitsItESaItEE8EvalTypeiPP11mxArray_tag12inDebugCheckP12_pcodeheaderPib+00002769
[ 35] 0x0000000107c7d7cd /Applications/MATLAB_R2014b.app/bin/maci64/libmwm_interpreter.dylib+01058765 _ZN12InterpBridge22EvalCmdWithLocalReturnERKSbItSt11char_traitsItESaItEEPibb+00000141
[ 36] 0x000000010728f11d /Applications/MATLAB_R2014b.app/bin/maci64/libmwbridge.dylib+00061725 _Z28evalCommandWithLongjmpSafetyRKSbItSt11char_traitsItESaItEE+00000093
[ 37] 0x000000010728f804 /Applications/MATLAB_R2014b.app/bin/maci64/libmwbridge.dylib+00063492 _Z8mnParserv+00000772
[ 38] 0x000000010225a832 /Applications/MATLAB_R2014b.app/bin/maci64/libmwmcr.dylib+00288818 _ZN11mcrInstance30mnParser_on_interpreter_threadEv+00000034
[ 39] 0x00000001022350d1 /Applications/MATLAB_R2014b.app/bin/maci64/libmwmcr.dylib+00135377 _ZN3mcr7runtime17InterpreterThread4Impl22TypedInvocationRequestIvE29doNotRethrowMxArrayExceptionsERKN5boost8functionIFvvEEE+00000033
[ 40] 0x0000000102235d89 /Applications/MATLAB_R2014b.app/bin/maci64/libmwmcr.dylib+00138633 _ZN5boost6detail11task_objectIvNS_3_bi6bind_tIvPFvRKNS_8functionIFvvEEEENS2_5list1INS2_5valueIS6_EEEEEEE6do_runEv+00000025
[ 41] 0x0000000102235586 /Applications/MATLAB_R2014b.app/bin/maci64/libmwmcr.dylib+00136582 _ZN5boost6detail9task_baseIvE3runEv+00000166
[ 42] 0x0000000102235377 /Applications/MATLAB_R2014b.app/bin/maci64/libmwmcr.dylib+00136055 _ZN3mcr7runtime17InterpreterThread4Impl22TypedInvocationRequestIvE3runEv+00000023
[ 43] 0x0000000102230f90 /Applications/MATLAB_R2014b.app/bin/maci64/libmwmcr.dylib+00118672 _ZN3mcr7runtime17InterpreterThread4Impl26invocation_request_handlerEl+00000032
[ 44] 0x0000000101d70005 /Applications/MATLAB_R2014b.app/bin/maci64/libmwservices.dylib+00196613 _ZN10eventqueue18UserEventQueueImpl5flushEv+00001205
[ 45] 0x000000010874da65 /Applications/MATLAB_R2014b.app/bin/maci64/libmwuix.dylib+00043621 _Z25isHandleUserEventCallbackPFvPvE+00000501
[ 46] 0x0000000101e3fa45 /Applications/MATLAB_R2014b.app/bin/maci64/libmwservices.dylib+01047109 _Z21svWS_ProcessCallbacksv+00016693
[ 47] 0x0000000101e3b152 /Applications/MATLAB_R2014b.app/bin/maci64/libmwservices.dylib+01028434 _Z25svWS_ProcessPendingEventsiib+00000834
[ 48] 0x0000000101e3ae5e /Applications/MATLAB_R2014b.app/bin/maci64/libmwservices.dylib+01027678 _Z25svWS_ProcessPendingEventsiib+00000078
[ 49] 0x000000010223118f /Applications/MATLAB_R2014b.app/bin/maci64/libmwmcr.dylib+00119183 _ZN3mcr7runtime17InterpreterThread4Impl14process_eventsERKN5boost10shared_ptrIS2_EE+00000095
[ 50] 0x000000010223149f /Applications/MATLAB_R2014b.app/bin/maci64/libmwmcr.dylib+00119967 _ZN3mcr7runtime17InterpreterThread4Impl3runERKN5boost10shared_ptrIS2_EEPNS2_12init_contextE+00000335
[ 51] 0x0000000102217d1b /Applications/MATLAB_R2014b.app/bin/maci64/libmwmcr.dylib+00015643 _Z26run_init_and_handle_eventsPv+00000059
[ 52] 0x00007fff9c01d2fc /usr/lib/system/libsystem_pthread.dylib+00013052 _pthread_body+00000131
[ 53] 0x00007fff9c01d279 /usr/lib/system/libsystem_pthread.dylib+00012921 _pthread_body+00000000
[ 54] 0x00007fff9c01b4b1 /usr/lib/system/libsystem_pthread.dylib+00005297 thread_start+00000013
This error was detected while a MEX-file was running. If the MEX-file
is not an official MathWorks function, please examine its source code
for errors. Please consult the External Interfaces Guide for information
on debugging MEX-files.
If this problem is reproducible, please submit a Service Request via:
http://www.mathworks.com/support/contact_us/
A technical support engineer might contact you with further information.
Thank you for your help.
</pre>
```
## History
##### Vadim Pisarevsky on 2015-04-27 11:32
```
- Category set to highgui-images
```
##### Vadim Pisarevsky on 2015-05-22 07:26
```
- Category changed from highgui-images to matlab bindings
```
| bug,auto-transferred,priority: normal,affected: 3.4,category: matlab bindings | low | Critical |
97,439,430 | opencv | Incomplete docs -- solvePnP in python | Transferred from http://code.opencv.org/issues/4221
```
|| James Youngquist on 2015-02-27 08:50
|| Priority: Normal
|| Affected: 2.4.9 (latest release)
|| Category: documentation
|| Tracker: Bug
|| Difficulty: Easy
|| PR:
|| Platform: x64 / Linux
```
## Incomplete docs -- solvePnP in python
```
Please add the following information specific to calling solvePnP from Python to the docs at
http://docs.opencv.org/modules/calib3d/doc/camera_calibration_and_3d_reconstruction.html#solvepnp :
* Numpy array slices won't work as input because solvePnP requires contiguous arrays (enforced by the assertion using cv::Mat::checkVector() around line 55 of modules/calib3d/src/solvepnp.cpp version 2.4.9)
* The P3P algorithm requires image points to be in an array of shape (N,1,2) due to its calling of cv::undistortPoints (around line 75 of modules/calib3d/src/solvepnp.cpp version 2.4.9) which requires 2-channel information.
* Thus, given some data D = np.array(...) where D.shape = (N,M), in order to use a subset of it as, e.g., imagePoints, one must effectively copy it into a new array: imagePoints = np.ascontiguousarray(D[:,:2]).reshape((N,1,2))
```
## History
##### Vadim Pisarevsky on 2015-04-27 12:42
```
- Category set to documentation
```
| bug,auto-transferred,priority: normal,category: documentation,affected: 2.4 | low | Critical |
97,439,474 | opencv | viz::Camera should be usable without vtk | Transferred from http://code.opencv.org/issues/4224
```
|| Pavel Rojtberg on 2015-03-01 12:47
|| Priority: Normal
|| Affected: None
|| Category: viz
|| Tracker: Feature
|| Difficulty: Easy
|| PR:
|| Platform: None / None
```
## viz::Camera should be usable without vtk
```
the code is also in other contexts as well like ARUCO or using OpenGL directly. A option would be to move it into core like Affine3 another one is to select a subset of viz that can be compiled without VTK. In types.cpp (where the implementation of viz::Camera is) only viz::Widget::load depends on VTK.
```
## History
##### Pavel Rojtberg on 2015-03-01 12:49
```
the code is -also- *useful* in other contexts as well like ARUCO or using OpenGL directly.
```
| auto-transferred,priority: normal,feature,category: viz | low | Minor |
Subsets and Splits