Research on personnel positioning accuracy optimization technology for complex electric power environment

Personnel location information is the core of production safety control, where existing methods such as BeiDou positioning, UWB positioning and other methods are difficult to meet the positioning needs of complex power production environment. This paper proposes a visual localization method based on the optimization of personnel recognition accuracy. For the accurate identification of personnel location in the complex electric power environment, this paper proposes a pedestrian detection method based on the fusion of multiple target detection algorithms by extending the Kalman filtering method, fusion of target detection and key point detection model, to improve the personnel identification and localization ability. It is also validated based on the data of the substation safety demonstration base environment. The conclusion of the study shows that the range and stability of target detection can be significantly improved based on the method proposed in this paper.