PCB defect detection technology based on deep learning
Printed circuit board(PCB)is a key component of electronic products.It is important to detect the PCB defect timely and accurately in the production process.There are some problems with the traditional detection methods,such as slow speed,high cost and low accuracy.Aiming at the PCB defect detection problem,a series of experiments were conducted based on YOLO detection algorithm to test the average accuracy,precision,recall,and FPS(frames per second)in the same experimental environment.Experimental results showed that YOLOv7 has a certain improvement in accuracy than YOLOv5,and YOLOv5 is faster than YOLOv7 in training and inference.In order to improve the performance,an improved algorithm based on YOLOv5 was proposed,the network structure of YOLOv5 was fused the convolutional block attention module(CBAM).Experiments verified that the average accuracy is improved by 7.40％,the accuracy and recall rate are also improved by 3.57％and 5.63％than those of theYOLOv5 respectively.
万方数据
维普
