Update app.py

app.py

@@ -13,15 +13,8 @@ def draw_lines(image, hough_lines, height):
     if hough_lines is None:
         return image
 
-    # Filter lines based on their position (closer to the vehicle)
-    filtered_lines = []
-    for line in hough_lines:
-        x1, y1, x2, y2 = line[0]
-        # Only keep lines where the lowest point (max of y1, y2) is within a certain range from the bottom
-        if max(y1, y2) > height * 0.6:  # Focus on the lower 40% of the image
-            filtered_lines.append(line)
+    filtered_lines = [line for line in hough_lines if max(line[0][1], line[0][3]) > height * 0.6]
 
-    # Draw the filtered lines
     for line in filtered_lines:
         x1, y1, x2, y2 = line[0]
         cv2.line(image, (x1, y1), (x2, y2), (0, 255, 0), 2)
@@ -29,12 +22,11 @@ def draw_lines(image, hough_lines, height):
 
 def process(img):
     height, width = img.shape[:2]
-    # Adjust ROI to focus on a smaller area closer to the vehicle
     roi_vertices = np.array([
-        (width * 0.1, height),           # Bottom-left
-        (width * 0.4, height * 0.6),    # Top-left
-        (width * 0.6, height * 0.6),    # Top-right
-        (width * 0.9, height)            # Bottom-right
+        (width * 0.1, height),
+        (width * 0.4, height * 0.6),
+        (width * 0.6, height * 0.6),
+        (width * 0.9, height)
     ], np.int32)
     
     gray_img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
@@ -42,20 +34,11 @@ def process(img):
     edges = cv2.Canny(blurred_img, 50, 150)
     roi_img = roi(edges, roi_vertices)
     
-    # Adjust Hough Transform parameters to detect shorter, closer lines
-    lines = cv2.HoughLinesP(
-        roi_img,
-        rho=1,
-        theta=np.pi / 180,
-        threshold=50,
-        minLineLength=30,
-        maxLineGap=50
-    )
+    lines = cv2.HoughLinesP(roi_img, 1, np.pi / 180, 50, minLineLength=30, maxLineGap=50)
     
     return draw_lines(img, lines, height) if lines is not None else img
 
 def lane_detection(video_path):
-    # Check if the input video path is valid
     if not os.path.exists(video_path):
         return "Error: Video file not found."
 
@@ -65,56 +48,34 @@ def lane_detection(video_path):
     
     frame_width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
     frame_height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
-    fps = int(cap.get(cv2.CAP_PROP_FPS))
+    fps = max(1, int(cap.get(cv2.CAP_PROP_FPS)))  # Avoid division by zero
     
-    # Create a temporary file for the processed video
     temp_file = tempfile.NamedTemporaryFile(delete=False, suffix=".mp4")
     output_path = temp_file.name
     temp_file.close()
     
-    # Use H.264 codec for better compatibility
-    try:
-        fourcc = cv2.VideoWriter_fourcc(*"avc1")
-    except:
-        fourcc = cv2.VideoWriter_fourcc(*"mp4v")
-    
+    fourcc = cv2.VideoWriter_fourcc(*"mp4v")  # Use mp4v for broader compatibility
     out = cv2.VideoWriter(output_path, fourcc, fps, (frame_width, frame_height))
     if not out.isOpened():
         return "Error: Could not create output video file."
     
-    # Process each frame
     while cap.isOpened():
         ret, frame = cap.read()
         if not ret:
             break
-        
         processed_frame = process(frame)
         out.write(processed_frame)
     
     cap.release()
     out.release()
     
-    # Verify the output file exists and has content
-    if not os.path.exists(output_path) or os.path.getsize(output_path) == 0:
-        return "Error: Output video file was not created or is empty."
-    
-    # Return only the processed video path
-    return output_path
+    return output_path if os.path.exists(output_path) and os.path.getsize(output_path) > 0 else "Error: Output video file was not created."
 
-# Use gr.Blocks instead of gr.Interface for better compatibility with Hugging Face Spaces
-with gr.Blocks(title="Road Lane Detection Demo") as demo:
-    gr.Markdown("## Road Lane Detection Demo")
-    gr.Markdown("Upload a video to detect road lanes. The app will display the processed video with detected lanes.")
-    
-    with gr.Row():
-        video_input = gr.Video(label="Original Video", interactive=True)
-        video_output = gr.Video(label="Processed Video")
-    
-    submit_btn = gr.Button("Submit")
-    submit_btn.click(
-        fn=lane_detection,
-        inputs=video_input,
-        outputs=video_output
-    )
+demo = gr.Interface(
+    fn=lane_detection,
+    inputs=gr.Video(label="Original Video"),
+    outputs=gr.Video(label="Processed Video"),
+    allow_flagging="never"
+)
 
-# No need to call demo.launch() in Hugging Face Spaces
+demo.launch()