Doesn't need one. This is not an NP-hard problem, nor is the proposed solution. It could, in fact, be done on a relatively low budget. Relative, that is, to the costs that might otherwise come with lawsuits or bad media publicity resulting from "unauthorized" video taken by passengers, in any event.
Besides, if he does the footwork and engineering on his customer's dime, he (and perhaps the customer, as a partner in a joint venture) could come out of it with a product they can market and sell to other similar bus operators, theaters, etc.
None of the tech I outlined is rare, odd, or terribly expensive (excepting LIDAR units, at about $5k each), and can be done entirely with COTS hardware and some customized programming that would only have to be built on top of programming that has already been done in this field. In fact, the PS3EYE or Kinect already has most of the capability needed.
Honestly, I've seen homebrewers do much more complex things (including code), such as the
home-built, automated air-hockey table built by a father for his daughter, which not only has to identify the puck, but then do loads of computational math to determine angles, speeds and force loads: what I initially described in my original post would almost be child's play by comparison.