> My understanding is that its too slow and not accurate enough for pinpoint tracking of targets (...)

I find these two assertions very surpsising. Which part of the process do you expect to be slow and what do you mean by slow? Why do you expect accuracy much smaller than the size of a pixel in an image?