I’ve been messing around with the Beta Kinect SDK and was wondering if there was any way to directly access the ir speckle field info that is returned to the kinect. I want to try to map a persons body (not just the skeleton) using triangulation of points on the body. I may be going about this the wrong way but I was thinking that since the Kinect is already processing info about thousands of dots on its target I could use a subset of these as my vertex set rather than generating the points myself.
Does anyone know if this is possible? I would prefer to use c# but would be willing to dust off my c++ skills (and learn a few more) if necessary.
There is no way, currently, to access the raw IR data using the SDK. And trust me when I say you probably don’t want to/need to. The IR pattern thrown by the light built into the Kinect is not a simple uniformly spaced pattern. A Google search for “Kinect IR Pattern” will show you that the pattern isn’t even perfectly rectangular.
What you should use is the depth map computed by the Kinect. It takes the input from the thousands of IR dots and converts it into an easy to use (albeit somewhat noisy) 640 x 480 (or 320 x 240, or 180 x 60…) image. This should suffice for mapping a person’s body, especially since there are methods in the SDK to translate between points on a skeleton and the corresponding points in the depth map.