I agree with squirrel; lowering the size of the pictures (which I believe we did last year) is the best way to reduce camera lag. In addition, any image processing being done (ie scanning the picture for a pattern/etc.) should be kept to a minimum. Do some tests, like constructing a model target and seeing how detailed it is from 60' away. If it seems really detailed, then cut down the image size. You don't need
a ton of detail for the target; a 2 or 3 pixel thick black oval will do. (Cutting down the image size reduces the amount of data that must travel, and the amount of data one must sift through if one is searching through the image for a target, and the amount of RAM used).
I haven't looked in to the Camera's documentation at all, but if it offers different functions for kinds of color scanning (for example, scanning for RGB values instead of HSV), try scanning for RGB; I have *heard* that RGB functions are slightly faster than HSV, but that was on a different platform (Pascal programming) and obviously a different compiler, so take this worth a grain of salt, if even that.
