This forum is disabled, please visit https://forum.opencv.org
 | 1 | initial version |
The Kinect should function with OpenNI; probably that's the best way to use it. If you use OpenNI, you can capture directly using OpenCV. But Libfreenect is also good.
On the other hand, as the document says, you don't have to calibrate the Kinect, only if you need some very accurate measurements. It already has a factory calibration.
To correct the offset between the RGB and depth images, you have to enable the RGB-depth registration with the freenect_map_rgb_to_depth function.
Otherwise it's not a good idea to do edge detection on the depth image directly. It has some "holes" (parts without depth data) which will always give strong contours.
