Abstract: The combination of edge-end devices and transmission line intelligent inspection can meet the needs of real-time defect detection in the field. However, the current research on algorithms for edge-end devices applicable to low-computing-power, low-memory devices is rarely available. Aiming at the above problems, this paper proposes a real-time defect detection method based on YOLO-GSS transmission line edge-end. Firstly, Mosaic-9 is used to improve the input end of YOLOv8 network, which improves the number of input features of the algorithm and enhances the robustness of the algorithm. Then, GhostNet and S-FPN are introduced to improve the Backbone and Neck part, which improves the inference speed of the algorithm and corrects the accuracy at the same time. Finally, SIoU is used to correct the YOLOv8's CIoU loss function to further improve the detection accuracy of the algorithm. The experimental results show that, compared with the original YOLOv8, the method proposed in this paper can be adopted to realize a quattuor increase in inference speed on Nvidia Jetson NX edge-end devices without too much decrease in accuracy, which can meet the demand for real-time detection of defects on transmission line sites.