Lightweight detection model FHCYOLO for wind turbine damage

Wind turbines working in harsh environments are very susceptible to environmental influences, which can cause some damages. In order to detect blade defects timely and effectively, a lightweight target detection method FHCYOLO is proposed in this study. The network structure of the model consists of lightweight modules FasterNet, HS-FPN, and CA Attention Module, which ensures the detection accuracy and reduces the model parameters at the same time. The lightweight level of the model is enhanced by integrating the lightweight structure with YOLO11. The detection accuracy of the model is enhanced by feature selection and enhancement of CA attention. The mAP50 and mAP50-95 of the model are improved by $0.9 \%$ and $0.6 \%$, respectively, the GFLOPs of the model are reduced by $38.1 \%$, and the FPS is improved by 6 frames per second.