PCB Defect Detection Method Based on Improved YOLOv11s

In the production process of printed circuit boards, defects such as opens, shorts, and mousebites are inevitable due to process imperfections, material issues, and environmental factors. At present, many existing PCB defect detection methods have shortcomings such as low detection accuracy and low detection efficiency. This paper proposes a PCB defect detection method based on improved YOLOv11s. On the basis of the original YOLOv11s model, a detection head is added for detecting small targets, and the convolutional block attention module is introduced to focus the model on key information in the image, and Inner IoU loss function is used to optimize the sensitivity of the model to target localization accuracy. The improved model achieves 97.90% mAP, a 2.59% increase over the original YOLOv11s’s 95.31%. With 95.76FPS detection speed, it enables real-time PCB defect detection through video.