Overhead transmission line robot can replace artificial power inspection tasks, solve the low efficiency in manual inspection Labor intensity The problem such as intelligent degree is insufficient Put forward between your checking design train of thought, through the design of mechanical structure optimization identification algorithm, realize the independent walking of the robot Wireless charging image acquisition intelligent inspection In the aspect of software, a special program and interface are developed based on C#, and a data set based on overhead transmission line hardware and defect is established. On the basis of YOLOv4 target detection algorithm, through data enhancement The boundary box optimization and model structure improvement, the establishment of YOLOV4-M model, optimize the performance of power hardware identification and detection, improve the environmental adaptability of the system application through the robot online operation and testing, the identification algorithm in the speed of 45fps, the average detection accuracy of 97.6%.