Research on Foreign Object Detection and Localization of Power Transmission Lines Based on CAFM-YOLO Model and Binocular Stereo Vision Technology

With the rapid growth of electricity demand, the safe and stable operation of transmission lines is crucial for the protection of the power grid. However, foreign objects invading transmission lines can easily cause faults and have a serious impact on the power system. Due to the low efficiency of traditional manual inspection methods, with the advancement of technology, intelligent inspection based on drones has gradually become mainstream. This paper proposes a target detection algorithm based on CAFM-YOLO, and combines binocular stereo vision technology to achieve accurate detection and localization of foreign objects in transmission lines. The experimental results show that the CAFM-YOLO model proposed in this paper outperforms existing mainstream object detection algorithms in terms of detection accuracy and speed, and the inference time can meet the real-time requirements of drone inspection. At the same time, this paper uses binocular vision to estimate depth maps and obtain the specific location of foreign objects, which will contribute to the automation and intelligent development of the field of transmission line inspection.