I posted here about the Z-Cam by 3DV Systems. It is a camera that uses time-of-flight from a set of LED's that emit a pulse. The special imaging chip that has a fast gate so that each pixel can have a depth value.
I imagine that this technology is going to revolutionize navigation for robots. The best part is that they want to market it to gamers, and price it under $100.
For computer vision, it can provide a method for creating a 3-d map of the environment. For your specific application, it provides a way to filter out objects in the background. I believe this "depth cropping" is onboard the camera, so it doesn't need computer processing to do it.
It is supposed to be out this year, but doesn't seem to be for sale yet to the public. They do have a SDK for sale (and I imagine the camera is available to developers as well).
Joe Dunfee