AML/V : An Industrial Machine Vision Programming System

Abstract

The use of robots rather than human workers to perform industrial tasks has a number of disadvantages. One is that parts and tools must be presented to the robot at ac curate positions and orientations. Another is that robots cannot do implicit in-process inspection the way people do. These two disadvantages can be reduced through the use of machine vision to determine the position, orienta tion, and even identity of parts and tools in the workspace and to look for qualitative and quantitative defects. This paper describes AML/V, an experimental hardware and software extension of an existing robot programming sys tem, intended for industrial machine vision programming.