Detection of surface defects of printed circuit board based on YOLOv5
An improved YOLOv5 detection model was proposed for the unsatisfactory detection effect of PCB surface defects.The shallow feature map was used to reunite the anchor frame according to the pixel size of PCB surface defects in the data set.The NGWD loss function was used instead of the CIoU localization loss function,and the Biformer attention mechanism was added.The average detection accuracy of 0.953 was achieved while ensuring the lightweight of the model.The improved algorithm could effectively improve the surface defect detection ability of printed circuit board.
万方数据
维普
