Standard Compliance Detection in Power Field Operations Based on Improved YOLOv7

In recent years, power field construction accidents have occurred frequently. The safety guarantee measures for construction personnel are thrown into sharp focus. Wearing safety harnesses is one of the most important requirements to protect the safety of construction personnel. Therefore, it is necessary to check whether personnel working at heights are wearing safety harnesses. The existing safety harness wearing detection algorithms face challenges such as target obstructions, substantial variance, and complex environments, leading to low detection accuracy. To address these issues, in this paper, a standard compliance detection model in power field operations based on improved YOLOv7 is proposed. The biformer attention mechanism is incorporated into the backbone network of YOLOv7 for the feature extraction, which can bolster the target detection performance. The experiments show that the detection precision of improved YOLOv7 is 4% higher than YOLOv7. Meanwhile, the mAP@0.5 and mAP@0.5:0.95 are 0.4% and 1.6% higher than YOLOv7, which meet the accuracy requirement of the safety risk management in the field operations of power systems.