Research on Multi-Class Component Detection of Transmission Lines Based on Improved YOLO11

As the backbone of power systems, transmission lines require regular inspections to promptly address potential safety hazards, which is essential for ensuring the stable operation of the power grid. In drone-based inspection tasks for transmission lines, the complex backgrounds in aerial images often lead to missed detections and false positives, a problem that is particularly pronounced in the detection of small targets. To address this limitation, we propose an improved YOLO11-based object detection algorithm. First, we design a novel C3k2-PMSC module, which effectively suppresses the interference of complex background information and enhances feature extraction capabilities. Second, we propose a lightweight and precisionbalanced small target detection module (LPBSTD), which strengthens the extraction of small target features and improves small target detection performance. Finally, we replace the original upsampling method with DySample, preserving rich image details and further enhancing small target detection capabilities. Experimental results show that, compared to the original network, the improved algorithm increases precision, recall, and mean average precision by 5.4%, 4.6%, and 6.3%, respectively, demonstrating the effectiveness of the proposed improvements.