WCC-YOLO: An improved YOLO11 model for defect detection of EL photovoltaic cells

With the increasing demand for new energy, the importance of photovoltaic systems as a sustainable energy solution is becoming increasingly prominent. However, defects such as cracks that may occur during the operation of photovoltaic cells can reduce power generation efficiency and cause safety hazards. This study introduces the WCC-YOLO model, an improved YOLO11-based approach for detecting defects in EL photovoltaic cells. By integrating WTConv, C6SA, C3Contmix, and ECA attention mechanisms, it enhances feature extraction and detection accuracy. Tested on the PVEL-AD dataset, WCC-YOLO achieves 90.3% accuracy, 89.7% recall, and 94.0% mAP, outperforming common YOLO and RTDETR models, with only 9.06M parameters. Ablation experiments and attention mechanism comparisons validate its component effectiveness. Result visualizations show superior performance in identifying small targets and complex backgrounds. WCC-YOLO offers an efficient and accurate solution for photovoltaic cell defect detection.