AUTOMATED VISUAL INSPECTION OF PRINTED CIRCUIT BOARDS

Abstract

In this project, an automated visual inspection system is built to inspect printed circuit boards for circuit line defects. The inspection algorithm employed is based on the principle of connectivity-preserved shrinking (CPS). The image of a sample printed circuit board is first scanned using a single-line camera and stored in memory, The image is initially processed using thresholding and filtering techniques to obtain a "clean" binary image for inspection. It is then inspected to detect defects in the circuit lines, The processing hardware consists of an Alpha microprocessor (running at 300 MHz) and a host PC. The entire process from printed circuit board scanning to output image takes about 8 seconds.