Update app.py
Browse files
app.py
CHANGED
@@ -1,22 +1,30 @@
|
|
1 |
import io
|
2 |
import time
|
|
|
3 |
import numpy as np
|
4 |
import cv2
|
5 |
-
import torch
|
6 |
-
from transformers import DPTFeatureExtractor, DPTForDepthEstimation
|
7 |
from fastapi import FastAPI, File, UploadFile
|
8 |
from PIL import Image
|
9 |
import uvicorn
|
|
|
10 |
|
|
|
11 |
app = FastAPI()
|
12 |
|
13 |
-
# 🟢
|
14 |
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
15 |
|
16 |
-
# 🟢 Tải model
|
17 |
-
|
18 |
-
|
19 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
20 |
|
21 |
@app.post("/analyze_path/")
|
22 |
async def analyze_path(file: UploadFile = File(...)):
|
@@ -24,27 +32,22 @@ async def analyze_path(file: UploadFile = File(...)):
|
|
24 |
image_bytes = await file.read()
|
25 |
image = Image.open(io.BytesIO(image_bytes)).convert("RGB")
|
26 |
|
27 |
-
# 🔵 Resize
|
28 |
-
|
29 |
-
|
30 |
-
# 🟢 Chuẩn bị ảnh cho mô hình
|
31 |
-
inputs = feature_extractor(images=image, return_tensors="pt").to(device)
|
32 |
|
33 |
-
# 🟢
|
34 |
start_time = time.time()
|
35 |
-
|
36 |
-
# 🟢 Dự đoán Depth Map với DPT-Hybrid
|
37 |
with torch.no_grad():
|
38 |
-
|
39 |
-
|
40 |
-
# 🟢 Xử lý ảnh sau khi dự đoán
|
41 |
-
predicted_depth = outputs.predicted_depth.squeeze().cpu().numpy()
|
42 |
-
depth_map = (predicted_depth * 255 / predicted_depth.max()).astype("uint8")
|
43 |
-
|
44 |
end_time = time.time()
|
45 |
-
print(f"⏳
|
|
|
|
|
|
|
|
|
|
|
46 |
|
47 |
-
# 🟢
|
48 |
start_detect_time = time.time()
|
49 |
command = detect_path(depth_map)
|
50 |
end_detect_time = time.time()
|
|
|
1 |
import io
|
2 |
import time
|
3 |
+
import torch
|
4 |
import numpy as np
|
5 |
import cv2
|
|
|
|
|
6 |
from fastapi import FastAPI, File, UploadFile
|
7 |
from PIL import Image
|
8 |
import uvicorn
|
9 |
+
from torchvision import transforms
|
10 |
|
11 |
+
# 🟢 Tạo FastAPI
|
12 |
app = FastAPI()
|
13 |
|
14 |
+
# 🟢 Kiểm tra GPU
|
15 |
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
16 |
|
17 |
+
# 🟢 Tải model MiDaS
|
18 |
+
midas = torch.hub.load("isl-org/MiDaS", "DPT_Swin2_L_384")
|
19 |
+
midas.to(device)
|
20 |
+
midas.eval()
|
21 |
+
|
22 |
+
# 🟢 Chuẩn bị bộ tiền xử lý ảnh
|
23 |
+
transform = transforms.Compose([
|
24 |
+
transforms.Resize((384, 384)),
|
25 |
+
transforms.ToTensor(),
|
26 |
+
transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
|
27 |
+
])
|
28 |
|
29 |
@app.post("/analyze_path/")
|
30 |
async def analyze_path(file: UploadFile = File(...)):
|
|
|
32 |
image_bytes = await file.read()
|
33 |
image = Image.open(io.BytesIO(image_bytes)).convert("RGB")
|
34 |
|
35 |
+
# 🔵 Resize và chuẩn hóa ảnh
|
36 |
+
input_tensor = transform(image).unsqueeze(0).to(device)
|
|
|
|
|
|
|
37 |
|
38 |
+
# 🟢 Dự đoán Depth Map với MiDaS
|
39 |
start_time = time.time()
|
|
|
|
|
40 |
with torch.no_grad():
|
41 |
+
depth_map = midas(input_tensor)
|
|
|
|
|
|
|
|
|
|
|
42 |
end_time = time.time()
|
43 |
+
print(f"⏳ MiDaS xử lý trong {end_time - start_time:.4f} giây")
|
44 |
+
|
45 |
+
# 🟢 Chuẩn hóa ảnh Depth Map
|
46 |
+
depth_map = depth_map.squeeze().cpu().numpy()
|
47 |
+
depth_map = (depth_map - depth_map.min()) / (depth_map.max() - depth_map.min()) * 255
|
48 |
+
depth_map = depth_map.astype("uint8")
|
49 |
|
50 |
+
# 🟢 Xử lý phát hiện đường đi
|
51 |
start_detect_time = time.time()
|
52 |
command = detect_path(depth_map)
|
53 |
end_detect_time = time.time()
|