An Improved Automatic Threshold Segmentation Method Used in PCBA Vision Inspection

摘要

Image segmentation is an important step in vision inspection of Printed Circuit Board Assembly (PCBA), which has effect on the validity of next work. Based on summarizing the previous results, several common methods for image segmentation are compared and analysed. According to the darker image colour and lower contrast of PCB, an automatic threshold segmentation method, which is based on mathematical morphology and statistical characters, is presented in the paper. The experiment results show that the segment effect and processing speed is better than the existing threshold image segmentation algorithms, and can meet the need of PCBA on-line visual inspection.