Detection and location technology of substation personnel based on EfficientDet

Abstract

Artificial intelligence technology has been widely used in the field of object detection and recognition. Limited by the complex working environment of power grid, the application of artificial intelligence in power grid system is just beginning. This paper focuses on a new neural network structure proposed by Google: EfficientDet. Based on the analysis of its characteristics, we applied the EfficientDet model to detect substation workers. The feasibility of applying EfficientDet model to substation personnel detection is verified. The advantages and disadvantages of EfficientDet model used in substation personnel detection and location are proved by experiments, and the possible improvement direction in the next step is discussed. The research results of this paper can also be applied to the environment of personnel location and detection in other power grid systems with a few changes.