Solar cell surface defect detection based on optimized YOLOv5

Traditional vision methods for solar cell defect detection have problems such as low accuracy and few types of detection, so this paper proposes an optimized YOLOv5 model more accurate comprehensive identification defects in cells. The firstly integrates five data enhancement methods, namely Mosaic, Mixup, hsv transform, scale transform flip, to expand the existing set improve feature training enhance robustness model; secondly, CA attention mechanism is introduced extraction ability address different target classification localization concerns, head original replaced with a decoupling head, which significantly without affecting convergence speed model. results show that achieves mAP 96.1% on publicly available dichotomous ELPV dataset, can identify locate variety common PVEL-AD while reach 87.4%, improvement 10.38% compared model, enables perform accurately ensuring real-time requirement surface task.