CGFM-YOLO: A Foreign Object Detection Algorithm for Transmission Lines Based on an Improved YOL011

Foreign object detection on transmission lines is crucial to ensure the safe operation of power systems. Aiming at the limitations of the existing YOL011 model in terms of foreign object detection accuracy, this paper proposes an improved method. The ContextGuideFM module is introduced to replace the original concat module, aiming to improve the feature pyramid network (FPN) in YOL011. The design of this module takes into account the guidance and adaptive adjustment of contextual information during the multi-scale feature fusion process. Meanwhile, the SPPF is improved using the LSKA attention mechanism to enhance the multi-scale feature extraction capability. The experimental results show that the average accuracy (mAP) of the improved model in the transmission line foreign object detection task is improved from 79.6% to 81.8%, an increase of 2.2 percentage points.