Fusing multiple Kinect BodyFrame data RRS feed

  • Question

  • Hello, 

    I am working in an environment with 6 Kinects, which have overlapping fields of view. I would like to fuse the data streams to get a global view of the scene. 

    For beginnings, it would be very helpful to be able to fuse the bodyFrames of multiple kinects. 

    Simplest setup: 

    two cameras with an overlapping FOV see one person in the shared volume. How to fuse the data from the two kinects (all networking stuff aside)?

    I see for instance that every person being tracked generates a trackingID, however this tracking ID is not universal, meaning that that ulong is not the same on different machines...

    Is the only way to get this to work to fuse all HD images correctly and process the result by OpenCV? Is there no other way?

    I also understand that this is a computer vision problem.

    Thanks for any help 


    Tuesday, July 28, 2015 9:51 AM