Using undistorted camera information for feature detection RRS feed

  • Question

  • Hi,

    Just a foreword - we were part of the Kinect2 early release program and officially the Kinect2 is an amazing piece of hardware, well done!

    About our issue - We would like to use the raw RGB camera data for feature detection, but to do this would need to undistort the image (remove barrel distortion). To do this a few parameters seem to be needed, specifically the radial and tangential distortion coefficients.

    It seems that something like this can be read directly from the sensor? If not, would you happen to have a reference with the values?


    Wednesday, December 10, 2014 6:11 PM