A low cost artificial intelligence vision system for piece part recognition and orientation

SUMMARY Traditional feature extraction vision techniques are well proven. However, in certain circumstances they suffer from some severe disadvantages such as processing time, image complexity, internal features and lighting levels. This paper describes the use of a hybrid system using artificial intelligence techniques in combination with traditional feature extraction techniques which overcome some disadvantages. This hybrid system is able to locate and orientate parts with complex internal features and to provide an indication of quality.