Deep Learning-Based Method for Detecting Foreign Object Intrusions in Power Transmission Lines Using Drone Aerial Photography

The issue of bird nests on power transmission lines has garnered significant attention in power systems due to the potential risk of equipment failure and large-scale blackouts. This paper proposes a bird nest detection method based on an improved YOLOv8 model. The method integrates the Convolutional Block Attention Module (CBAM) and Atrous Spatial Pyramid Pooling (ASPP) into the YOLOv8 model to enhance the model's ability to capture bird nest features and improve detection accuracy. CBAM enhances the model's sensitivity to target features by combining spatial and channel attention mechanisms, while ASPP improves the model's detection capabilities across different scales through multi-scale feature extraction. Extensive experiments demonstrate that the improved YOLOv8-based bird nest detection method excels in terms of accuracy, recall, and detection speed. The application of this method is expected to significantly improve the efficiency and reliability of bird nest detection during power line inspections, thereby ensuring the stable operation of power systems.