Composite insulator identification and monocular ranging method based on YOLOv9

Aiming at the task of UAV inspecting composite insulators in transmission lines, this paper proposes a method of composite insulator target recognition and monocular distance measurement based on yolov9 algorithm. With the help of visible camera, the auxiliary distance measurement function is realized, which provides strong technical support for UAV to fly safely, take as clear photos as possible, and improve the line operation and maintenance ability. Firstly, the composite insulator pictures of various voltage levels are collected to make a data set, and the data set is expanded through the data enhancement strategy to improve the robustness of composite insulator detection. Then, yolov9 algorithm is used to carry out in-depth learning and training on the data set, and the detection accuracy of composite insulators at various voltage levels is 92.7%. Finally, the monocular ranging model of composite insulator is established based on the monocular ranging principle, and the ranging correction model is designed for the ranging error caused by pitch angle. The experimental results show that the modified model reduces the ranging error from 17.6% to 5.2% when the pitch angle is 30°; The influence of different weather conditions on the ranging accuracy is verified by experiments. If the brightness is too strong or too weak, the ranging error will increase. The ranging accuracy under strong light conditions is lower than that under weak light conditions. When the brightness adjustment reaches+100%, the ranging error will reach 11.67%.