Research and Implementation of PCB Defect Detection Based on Improved YOLOv5 Algorithm

Aiming at the problems of low detection accuracy and inaccurate defect location of PCB defects in conventional detection algorithms, an improved YOLOv5s PCB defect detection algorithm is proposed by studying PCB defect detection algorithms. Firstly, the loss function is optimized, and the original regression loss calculation method is changed from GIOU to EIOU, so as to accelerate the convergence speed of the model and improve the positioning accuracy of the frame, and the confidence loss balance coefficient is adaptively adjusted to improve the training effect of the model. Secondly, the attention mechanism of ECANet is introduced to improve the saliency and feature expression ability of PCB defect target area, and to enhance the extraction of small target defect detail information which is difficult to detect. Finally, in order to further verify the effectiveness and rationality of the improved YOLOv5s algorithm, ablation experiments are used for comparative analysis. The experimental results show that the improved YOLOv5s algorithm improves the detection accuracy (mAP) to 96.80%, which meets the requirements of PCB defect detection.