Method for Automatic Visual Defect Detection in Printed Circuit Boards

摘要

The report presents the results of an investigation of printed circuit board (PCB) inner layer visual inspection. A new algorithm is presented for comparison of detailed images. Results of the search for several types of defects in simulated PCB images under conditions of image-to-image misregistration, intensity differences, and noise are presented. These results indicate the algorithm can detect a wide variety of small defects. The requirements for components of a visual inspection system have been examined, including the optical detector head, and data processing hardware.