Abstract:Power grid operation involves high-voltage power equipment, point equipment and complex operation scenarios, personnel are faced with electric shock, electric shock, high temperature and other dangers, and there are serious personal safety risks. Traditional personal risk perception is limited by manual intervention, with problems such as insufficient perception ability and poor timeliness. Manual operation is also prone to subjective judgment and fatigue and other factors, leading to risk omission or misjudgment. In addition, the current prevention and control technology is mainly concentrated in a single dimension, relying on manual monitoring, low efficiency, high cost, and lack of intelligent, three-dimensional, all-round comprehensive prevention and control means. In view of the above problems, this paper intends to use advanced perception technology, image processing technology, artificial intelligence and other methods to put forward the two-stage behavior dynamic risk identification method of complex operators based on image screening, so as to realize the accurate perception, timely warning and three-dimensional prevention and control of personal risks in complex working environment. By studying key technologies such as complex operation characteristic risk elements, live equipment identification, and personnel behavior analysis, it will provide more reliable and efficient personal risk management and prevention and control means for the power grid industry, improve the level of operation safety, and reduce the risk of casualties. .