|
.. _object_segmentation: |
|
|
|
Object Segmentation |
|
=================== |
|
|
|
.. _3d_object_segmentation: |
|
|
|
3D object segmentation |
|
------------------------- |
|
|
|
Our scene segmentation works with the assumption that the table has to be parallel |
|
to the ground as we use |
|
`SampleConsensusModelPerpendicularPlane <https://pointclouds.org/documentation/classpcl_1_1_sample_consensus_model_perpendicular_plane.html>`_ |
|
from PCL library. We keep points which are perpendicular to `z` axis. |
|
|
|
There are two packages you can launch in order to segment table top point clouds: |
|
|
|
*mir_object_segmentation* package |
|
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ |
|
|
|
`mir_object_segmentation <https://github.com/b-it-bots/mas_industrial_robotics/tree/melodic/mir_perception/mir_object_segmentation>`_ |
|
package can be used to accumulate clouds from different views (if necessary), |
|
register them, and then segment the objects above the table. |
|
|
|
**Usage** |
|
|
|
* Find plane height |
|
|
|
To find the object height, you have to send the following messages to `event_in` |
|
topic: |
|
|
|
.. code-block:: bash |
|
|
|
rostopic pub /mir_perception/scene_segmentation/event_in std_msgs/String e_start |
|
rostopic pub /mir_perception/scene_segmentation/event_in std_msgs/String e_add_cloud_start |
|
rostopic pub /mir_perception/scene_segmentation/event_in std_msgs/String e_find_plane |
|
|
|
The plane height will be published to `workspace_height` topic: |
|
|
|
.. code-block:: bash |
|
|
|
/mir_perception/scene_segmentation/output/workspace_height |
|
|
|
|
|
* Segment objects from one view point |
|
|
|
.. code-block:: bash |
|
|
|
rostopic pub /mir_perception/scene_segmentation/event_in std_msgs/String e_start |
|
rostopic pub /mir_perception/scene_segmentation/event_in std_msgs/String e_add_cloud_start |
|
rostopic pub /mir_perception/scene_segmentation/event_in std_msgs/String e_segment |
|
rostopic pub /mir_perception/scene_segmentation/event_in std_msgs/String e_stop |
|
|
|
* Segment objects from one view point |
|
|
|
You can segment table top objects from multiple viewpoints. You can do this by |
|
manually moving the camera position to a new view point, and then register |
|
the cloud with the existing ones. |
|
|
|
.. code-block:: bash |
|
|
|
|
|
rostopic pub /mir_perception/scene_segmentation/event_in std_msgs/String e_start |
|
rostopic pub /mir_perception/scene_segmentation/event_in std_msgs/String e_add_cloud_start |
|
rostopic pub /mir_perception/scene_segmentation/event_in std_msgs/String e_add_cloud_stop |
|
|
|
|
|
rostopic pub /mir_perception/scene_segmentation/event_in std_msgs/String e_add_cloud_start |
|
rostopic pub /mir_perception/scene_segmentation/event_in std_msgs/String e_add_cloud_stop |
|
|
|
|
|
rostopic pub /mir_perception/scene_segmentation/event_in std_msgs/String e_segment |
|
rostopic pub /mir_perception/scene_segmentation/event_in std_msgs/String e_stop |
|
|
|
|
|
.. hint:: Published msgs from mir_object_segmentation |
|
|
|
* Object list |
|
|
|
.. code-block:: bash |
|
|
|
/mir_perception/scene_segmentation/output/object_list |
|
|
|
* Bounding Boxes (for visualization in Rviz) |
|
|
|
.. code-block:: bash |
|
|
|
/mir_perception/scene_segmentation/output/bounding_boxes |
|
|
|
* Object labels (for visualization in Rviz) |
|
|
|
.. code-block:: bash |
|
|
|
/mir_perception/scene_segmentation/output/labels |
|
|
|
* Debug pointcloud (shows filtered input to plane segmentation) |
|
|
|
.. code-block:: bash |
|
|
|
/mir_perception/scene_segmentation/output/debug_cloud |
|
|
|
* Object clusters: segmented point clouds |
|
|
|
.. code-block:: bash |
|
|
|
/mir_perception/scene_segmentation/output/tabletop_clusters |
|
|
|
* Workspace height: |
|
|
|
.. code-block:: bash |
|
|
|
/mir_perception/scene_segmentation/output/workspace_height |
|
|
|
|
|
*mir_object_recognition* package |
|
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ |
|
|
|
`mir_object_recognition` uses the same `scene_segmentation` method |
|
as in `mir_object_segmentation`, but it does not accumulate point clouds. It only |
|
takes one view and then segment table top objects. |
|
|
|
**Usage** |
|
|
|
* Segment objects |
|
|
|
.. code-block:: bash |
|
|
|
rostopic pub /mir_perception/multimodal_object_recognition/event_in std_msgs/String e_start |
|
|
|
**Published messages** |
|
|
|
.. code-block:: bash |
|
|
|
/mcr_perception/object_detector/object_list |
|
/mir_perception/multimodal_object_recognition/output/bounding_boxes |
|
/mir_perception/multimodal_object_recognition/output/debug_cloud_plane |
|
/mir_perception/multimodal_object_recognition/output/pc_labels |
|
/mir_perception/multimodal_object_recognition/output/pc_object_pose_array |
|
/mir_perception/multimodal_object_recognition/output/rgb_labels |
|
/mir_perception/multimodal_object_recognition/output/rgb_object_pose_array |
|
/mir_perception/multimodal_object_recognition/output/tabletop_cluster_pc |
|
/mir_perception/multimodal_object_recognition/output/tabletop_cluster_rgb |
|
/mir_perception/multimodal_object_recognition/output/workspace_height |
|
|
