Obviously, the first part of computer vision is obtaining visual imagery -- in effect, seeing. The image has to be translated into digital form, then processed by the computer. Most computer vision applications today use video cameras of some sort, but some "see" through different methods.