基于改进YOLOv5的锂电池极片缺陷检测方法

冉庆东,郑力新

Defect detection method of lithium battery electrode based on improved YOLOv5

Qingdong RAN,Lixin ZHENG

表 5 不同算法在锂电池极片数据集中的实验结果对比

Tab.5 Comparative experimental results of different algorithms on lithium battery electrode dataset

方法 P/106 t/ ms mAP50/ % mAP50S/ % mAP50L/ % mAP50∶95/ % Swin-Transformer[30] 37. 03 56.8 46.4 37.1 57.7 22.3 RetinaNet[5] 36.31 44.6 53.4 46.9 61.1 25.4 文献[10] 9.33 23.5 52.2 63.9 38.2 27.3 YOLOv5s 7. 04 22. 0 70.9 73.4 67.9 36.3 文献[11] 7. 08 24.1 70.4 72. 0 68.5 37.4 DDCNet-YOLOs (本研究) 6.53 22.8 72.4 74.4 70. 0 45.2 YOLOv5m 20.89 27.6 73.4 72.6 74.3 42.6 DDCNet-YOLO (本研究) 22.76 26.5 77.1 75.6 79. 0 45.7