Home Forums AR Sandbox Forum KinectViewer


Viewing 3 posts - 1 through 3 (of 3 total)
  • Author
  • #114688


    Has anyone had problems running the KinectViewer app? I have installed the AR Sandbox on about four different Linux systems. Two Debian based and two Mint based. Some use Vrui-4.4 and some Vrui-4.5. I don’t get any compile errors and all apps I try work fine except KinectViewer. For example, the world spins, RawKinectViewer works, even the Sandbox app works pretty well though I can’t finish the calibration without KinectViewer which just silently exits.

    Any ideas?

    Oliver Kreylos

    KinectViewer is not required for AR Sandbox calibration. Please check the complete installation instructions on this forum.

    Either way, KinectViewer’s minimal command line is

    KinectViewer -c 0

    to show 3D video from the first connected Kinect camera.


    Thanks for the info Oliver, yes I eventually figured that out. The confusing part is RawKinectViewer will accept:

    RawKinectViewer -c0
    RawKinectViewer -c 0
    and of course:

    KinectViewer -c0
    will silently fail/exit with no errors, and several other command line arguments, or switches will also cause it to exit silently, leaving one to wonder what happened if anything.

    We had some instructions from your site that I’m not able to find right now. They used KinectViewer to get the for corners of the sandbox. I believe that was step 6, and step 7 was the final calibration using CalibrateProjector. This worked fine after I learned to put a space between the c and 0. 🙂

    Say, CalibrateProjector seems to have a new feature where it draws a yellow box inside the sandbox. This box is quite a bit smaller than the boundaries of the sandbox entered in the previous step, and all the required tie points are within that smaller box. Soooo, we were wondering if you would consider a feature where we can pick the tie points with the mouse after the first pass is done. We would use this feature to get the edges of the sandbox not covered by the current set of tie points. Because when we check the results with the target disk on the wire, it starts diverging a bit when you get to the actual edges of the box. 🙂

    By the way, thanks for still being there and continuing to support this great project. That’s a lot of work I know.

    Clif Cox

Viewing 3 posts - 1 through 3 (of 3 total)
  • You must be logged in to reply to this topic.

Comments are closed.