Automatic visual inspection of printed circuit boards

The goal of the project is to develop effective and efficient methods that verify the integrity of printed circuit boards (PCBs) through the use of computer analysis of images. A system for using gray-scale and color images to detect and classify instances of part integrity defects was developed; the use of color imagery to identify color-coded resistors was investigated; and a system for verifying the identity of character-coded dual in-line package markings was developed. Current work in the area of part integrity involves the use of range data acquired with a structured light device. A prototype of a low-cost laser scanner for solder joint imaging has been designed and constructed. A graphical simulator of a much higher performance device was designed. By systematically varying parameters, the conditions under which various kinds of defects become apparent can be analyzed.