A robotic vision system with incorporation of artificial intelligence for industrial application

Abstract

Our main objective in this project is to develop a robotic vision system which is capable of inspecting IC and wafers for real-time application in an industrial environment. We have achieved this objective satisfactorily. IC's inspection rate can easily be maintained at 72 00 units per hour. This compares favourably with what is currently achieved by human inspectors. The inspection system for IC is then modified to inspect for die-related defects on wafers. We have achieved a performance reliability of over 90% for the wafer inspection system. We have also incorporated artificial intelligence in our robotic vision inspection system. This makes our system a bit slower due to longer processing time but more flexible and reliable. Using AI, more complicated inspection criteria can be included without significantly altering the software.