Intelligent Detection Method for Avian Nests on Transmission Lines Based on Edge Computing and YOLOv5

The safety of transmission lines is a prerequisite for the secure and stable operation of the power grid. Avian nests can severely affect the safety of overhead power transmission lines. Considering the issues such as the large number of parameters in existing avian nest detection and tracking algorithms, relatively complex networks, and high computational requirements, which are not conducive to deployment on embedded edge nodes, this paper proposes a lightweight avian nest detection model based on YOLOv5. Experimental results show that YOLOv5 can rapidly and accurately detect avian nests on transmission lines, achieving an accuracy of up to 90.15% and a detection speed of 34.2 frames per second. Compared with commonly used detection algorithms such as Faster-RCNN, SSD, YOLOv3, and YOLOv4, the YOLOv5 demonstrates stronger competitiveness in terms of avian nest recognition accuracy and speed. Therefore, the YOLOv5s algorithm ensures real-time detection while enhancing the detection accuracy of ambiguous fault target images, meeting the demand for unmanned aerial vehicles equipped with edge devices to conduct avian nest inspections on transmission lines.