Clothing violation detection under occlusion conditions

In order to quickly detect improper dressing of workers in the power industry and prevent accidents, this study proposes an improved YOLOv8n method. The method combines the CoordAtt attention mechanism and the small target detection layer. First, the CoordAtt attention mechanism is integrated into the backbone network to weight important regions and improve the model's attention to inappropriate clothing. Second, a specialized small target detection layer is introduced to improve the detection performance for small targets. By adjusting the network structure and parameters, the model can better capture the details of small targets, thus improving the detection accuracy and robustness. The experimental results show that the enhanced YOLOv8n model has significantly improved its performance in detecting inappropriate clothing worn by workers in the power industry. The accuracy reaches 89%, which is an improvement of 2.1% compared to the baseline model, and the mAP50 is improved by 2.5%. This improvement effectively increases the accuracy and robustness of the detection.