Improved the PCB Defect Detection Method Based on YOLOv5s

To address the challenges of weak anti-interference capability, limited feature expression ability, and high deployment difficulty in the defect detection of printed circuit boards (PCBs), we propose an enhanced PCB defect detection algorithm based on YOLOv5s. The C2f-SCSA module replaces the C3 module in the backbone network to enhance the multi-scale feature expression capability. Additionally, the EMAELAN4 module substitutes the downsampling module in the backbone network to reduce information loss during the downsampling process. The SPAA-HSFPN structure is utilized in the neck network to enhance multi-scale feature fusion capability. The proposed algorithm model achieves a mean average precision of 92.3% on the PCB dataset after data augmentation, representing a 3.0% improvement over the baseline model, while simultaneously reducing the model parameters by 17.2%.