none
Ask about kinect fusion for AR RRS feed

  • Question

  • Hi, I want to create an AR application using Kinect Fusion. 

    I want to insert a 3d model into my kinect fusion reconstruction. The expected output is like these two videos:

    https://www.youtube.com/watch?v=gChlRebNloA

    https://www.youtube.com/watch?v=MvP4cHfUD5g

    How to overlay the 3D object onto the 3d reconstruction?

    Is there function in the SDK that can be used to achieve my goal?

    Thanks

    ps: I use C#

    • Edited by Wahyu Setyo B Tuesday, March 25, 2014 1:26 PM additional detail
    Tuesday, March 25, 2014 1:22 PM

All replies

  • There is no support for injecting 3D traditional models into a Kinect Fusion pipeline. It can only process color/depth data to integrate into it reconstruction volume.

    From the videos, I would have to speculate, they are taking the 3D constructed volume data from Fusion and using that within their own AR space. There are certainly API's provided by Fusion to provide world coordinates based on the camera position. Determining surfaces and normal, would be a matter of 3D techniques, but that would not be something from the Kinect API's.

    Have a look at the Kinect Fusion Basics samples to understand how the integration pipeline works. You will have to break it down into what data you needs from the reconstructed world to determine the export data you need for your own AR space.

    http://msdn.microsoft.com/en-us/library/dn188670.aspx

    http://msdn.microsoft.com/en-us/library/dn188700.aspx


    Carmine Sirignano - MSFT

    Wednesday, March 26, 2014 9:13 PM