Comparison of Object Region Segmentation Algorithms of PCB Defect Detection

Abstract

As a core component of electronic products in industrial production, the printed circuit board (PCB) is highly integrated, and carries various electronic components and complex wire layout. Although the PCB has a small size, its defect detection directly affects the quality of circuit board, which is of great significance. This research aimed to study PCB defect detection based on machine vision technology, because the product quality inspection requirements have been continuously increasing in industrial modernization. Whether the object region segmentation algorithms are fast, effective, and accurate directly affects the effects and efficiency of subsequent machine vision defect detection, because object region segmentation is a key step in PCB defect detection. Three types of object region segmentation algorithms, namely, color space threshold segmentation, morphological edge detection segmentation, and K-means clustering segmentation, were studied, and their advantages and disadvantages were analyzed in detail. A suitable algorithm was selected for detection object through experiments, which laid the foundation for better algorithm improvement and segmented object regions quickly and accurately in the defect detection process.