PCB Defect Detection Algorithm Based on Lightweight YOLOv5
A printed circuit board(PCB)defect detection algorithm based on YOLOv5l improvement was proposed using Python on Windows to address the drawbacks of large network models and low accuracy in PCB defect detection.Six common defects were detected as datasets.Use the lightweight network EfficientNetLite0 as the backbone network of the model,and add P2 detection head to the feature pyramid to obtain smaller target features.The experimental results indicate that the proposed algorithm has the characteristics of high recognition accuracy,small model size,and fast detection for defects in printed circuit boards;the detection speed of a single image reaches 43.6 ms,the model size is 49.1 MB,and the accuracy indices of all categories reach 98.9％.The proposed algorithm provides a new idea for future industrial defect detection on edge device.
万方数据
维普
