Geometric Pattern Matching for Industrial Robot Guidance

Abstract

In spite of decades of work on object recognition in both the academic and industrial communities, the majority of industrial robotic applications are not vision guided, relying instead on mechanical fixturing and dead reckoning. The primary reason for this is that object recognition by machine vision simply has not worked well enough to be competitive. Recent developments in object recognition algorithms, driven by increasing demand for flexible automation and enabled by processor architectures well-suited to image analysis, have begun to change this picture. We will discuss requirements that must be satisfied for a method to achieve widespread use, trace the development of industrial object recognition, describe the present method, and mention some applications. Throughout we offer industrial perspective and experience to an academic forum, in hopes of better understanding.