Although I don't know all the specifics, our team got a 12-LED (or about that) LED ring and attached it around the camera lens with some double-sided foam tape.
I'm really not sure what you mean by "Axis camera servos." The cameras have mounts, and no motorized parts. Theoretically you could mount one on a servo and have a variable angle camera, but it's not an inherent part of the design.
I would not suggest a moving camera, if only because calculations that would normally have constants would have variables that would need to be updated or retrieved somehow, which means more complexity.
Cameras don't inherently track targets - they are stupid, and can only provide images. You the programmer are responsible for making the cRIO be the smarts. Your issue with the camera test crashing sounds like a memory leak (not freeing memory it should be). I ran into an almost identical issue with my tracking code during early testing.
I explained the theory behind tracking targets with the camera fairly extensively in
this thread, and once I have the new, simplified version of my code tested (which will happen tomorrow), I will be posting it so that all may benefit.