[~Cory~]

We use a negative image from the camera to find the black/white boundary. Then we compute possible polygons and put those into a genetic sorting algorithm that chooses the most likely target. The Heurisitics need a little more tunning but we can capture distance with a +- 2 inch accuracy.

In a more technical setting we are using a threaded socket connection from the DS computer to the robot. All vision processing happens offboard and is then sent back. We have used the system in competition without a hitch.