How to get raw data in uint16 format

[gustn6591]

Hello, I managed to get a grayscale image of (480,640) through the code below.

cap_depth = cv2.VideoCapture("rtsp://admin:admin@192.168.1.10/depth", cv2.CAP_GSTREAMER) while(True): ret2, frame_depth = cap_depth.read() cv2.namedWindow('kinova_depth', cv2.WINDOW_AUTOSIZE) cv2.imshow('kinova_depth',frame_depth) if cv2.waitKey(20) & 0xFF == ord('q'): break cap_depth.release() cv2.destroyAllWindows()

However, the output image was in uint8 format ranging from 0 to 255, and accurate depth information could not be obtained. I was able to get a value similar to uint16 through (pixel/255)*65535, but a lot of errors occurred because I had already brought the value of uint8. I did a lot of searching and found several ways, but I didn't know the exact way to get depth data in uint16 format from realsense(D410) attached to gen3. I know that using ros_rgbd.launch I can get the point cloud, so I'm pretty sure I can get uint16's information, but I don't know how. I want to know how to get depth information from ros_kortex_vision repo or an example, so I leave a question.

Thank you

[felixmaisonneuve]

Hi @gustn6591,

Depth stream is in 16bit encoding and we pass GRAY16_LE encoding to Gstreamer to get this point cloud. https://github.com/Kinovarobotics/ros_kortex_vision/blob/cd49bab1aa887b4791ac4d512ef181fd39361426/src/vision.cpp#L62-L66

https://github.com/Kinovarobotics/ros_kortex_vision/blob/cd49bab1aa887b4791ac4d512ef181fd39361426/src/vision.cpp#L120-L123

The color stream however is in RGB8 encoding https://github.com/Kinovarobotics/ros_kortex_vision/blob/cd49bab1aa887b4791ac4d512ef181fd39361426/src/vision.cpp#L124-L127

if you launch the kinova_vision_rgbd.launch file, you will only have access to both color and depth stream, so make sure you are using the correct depth stream and not the color stream.

Best, Felix