Object Detection Technology of Substation Patrol Robot Based on mYOLO Algorithms

In this paper, a m-yolo algorithm is proposed, which combines multiple groups of camera image data to realize the fast recognition and accurate positioning of the target in the substation. Inspection of equipment in substation mainly depends on manual work, which is labor intensive, inefficient and prone to negligence. Substation inspection robot has unique advantages in inspecting equipment in substation. Rapid detection and accurate identification of targets are the basis of normal operation of substation inspection robot. Based on the deep neural network, this paper realizes the YOLO detection algorithm in substation scenario, which provides technical support for the smooth operation of substation inspection robot.