LHF-YOLO: A Lightweight Network for the Detection of Safety Equipment for Working at Height

To address the challenge of real-time detection of safety equipment usage during work at height, this study introduces a lightweight detection model-LHF-YOLO. The backbone of the model employs a lightweight architecture that primarily comprises DWConv and an in-house developed C2fDGhost module. In the neck network, conventional convolution modules have been replaced with GSConv modules that incorporate shuffle operations. The detection head utilizes the YOLOv11 design. Finally, the Wise-IoUv2 loss function is applied to improve the accuracy of small object detection. The enhanced model achieves a 1.8% increase in [email protected] while reducing computational load and complexity by 41.2% and 42.7%, respectively. Consequently, the proposed method is both precise and efficient in detecting safety equipment for high-altitude operations.