Research on Real-time Analysis of Safety Behavior of Thermal Power Plant Operators Based on Video Data
Thermal power plant is a high-risk environment.Real-time analysis of operator safety behavior is very important to ensure personnel safety and the operation of thermal power plant.To this end,a real-time safety behavior analysis system for thermal power plant operators is designed for early warning of safety behavior such as whether the operators smoke,wear safety helmets,and whether they stay in dangerous areas.Aiming at the problem that the safety behavior detection method of power plant operators is susceptible to environmental interference,and there are problems of missed detection and false detection,this paper improves the YOLOv5 algorithm.Specifically,when YOLOv5 is used for feature fusion,the attention gating unit is used,which can significantly improve the safety behavior detection ability,and this module has fewer parameters to meet real-time requirements.Finally,the pro posed method was verified on the SHWD and smoking behavior datasets.The experimental results show that this method can significantly improve the ability of safety behavior detection,and can be applied to the real-time safety behavior analysis system of power plant operators.
thermal power plantsafety behavior analysisYOLOv5attention gated unit
NETL
NSTL
万方数据
