基于改进YOLOv5的冲压件缺陷检测方法研究

冲压件在生产过程中容易出现裂纹、划痕、起皱、凹凸点等缺陷。目前，生产线上的冲压件缺陷检测以人工检测为主，效率低，且容易造成漏检。为此，提出了一种基于改进YOLOv5模型的缺陷检测算法。为了提高缺陷部分的关注度，更好地聚焦缺陷，本文在YOLOv5模型的主干网络中引入CA注意力模块。为了进一步提升模型的精度，本文通过对比实验，将目标框损失函数改为 GIoU，提升了定位精度。实验表明，相较于原模型，改进后的YOLOv5模型精准度、召回率、mAP值均得到提升。

Research on Defect Detection Method of Stamping Parts Based on Improved YOLOv5

Stamped parts are prone to cracks, scratches, wrinkles, bumps and other defects in the production process. At present, the defect detection of stamped parts on the production line is based on manual detection, which is inefficient and prone to leakage. For this reason, a defect detection algorithm based on the improved YOLOv5 model is proposed. In order to improve the attention of the defective part and better focus the defects, this paper introduces the CA attention module in the backbone network of the YOLOv5model. To further improve the accuracy of the model, this paper improves the localization accuracy by changing the target frame loss function to GIoU through comparative experiments. The experiments show that compared with the original model, the improvedYOLOv5 model precision, recall, and mAP value are all improved.