Abstract
This paper describes the development of a microcomputer-based, low-cost binary vision system to be used in automatic applications such as the control of an industrial robot. The vision system processes binary images and recognizes objects based on geometric features such as the number of holes and the moments of invariant. The system is trained by showing it the different stable positions. The image processing software is written in 8088 Assembly code and can recognize up to 40 parts per minute.