locked
Does the Kinect Fusion engine work on distorted depth images? RRS feed

  • Question

  • I have been looking through the KinectFusionExplorerWPF code and I don't see any mention of CameraSpacePoint and hence no mapping of distorted depth image to undistorted camera space. It looks like the Fusion library consumes depth images directly in ProcessDepthData() where depth frames get converted to FusionFloatImageFrame using DepthToDepthFloatFrame() function. Does this mean that provided Fusion samples don't account for radial distortions seen in the Kinect IR images?
    Friday, December 12, 2014 9:52 PM

Answers

  • Please take a look at the native samples for an example of how to correct for radial distortion. It appears that this correction did not make it across to the WPF sample (though it should have)

    Chimera Scorn - Chris White

    Monday, December 15, 2014 7:37 PM

All replies

  • Please take a look at the native samples for an example of how to correct for radial distortion. It appears that this correction did not make it across to the WPF sample (though it should have)

    Chimera Scorn - Chris White

    Monday, December 15, 2014 7:37 PM
  • Thanks! I hope the corrections will be included in the WPF sample in the future. 
    Wednesday, December 17, 2014 6:59 AM