Does anybody know what are the correct values to setup the projection and view matrix for a XNA rendering in order to match with the Kinect RGB camera?
For example, if the skeleton head joint coordinates are (X,Y,Z) then if I draw a sphere at these coordinates in my 3D space, the rendering should overlap the head in the RGB image.
In the end I took a trial and error approach and determined the parameters. You can find more details here: http://www.mindtreatstudios.com/our-projects/kinect-3d-view-projection-matrix-rgb-camera/
I'm sure this is not the most scientific way of reconstructing the 3D space, but given the time-frame I had available and the needs of my project, they work well and I'm happy with the results.
