An Improved YOLOv11 Model for Detecting the Metal Roofing Tiles alongside the Railways

Metal roofing tiles along railways pose significant hazards, potentially causing service disruptions and accidents. However, manual inspections are inefficient, limited in coverage, and lack automated data recording, rendering them insufficient to meet escalating safety requirements. This study introduces an enhanced YOLOv11-based intelligent detection method aimed at overcoming the challenges posed by complex backgrounds, occlusion, and real-time performance in the detection of metal roofing tiles along railway lines. The proposed method integrates attention modules, dynamic snake convolutions, and an enhanced loss function, achieving significant improvements in detection accuracy and robustness. Experimental results on the railway dataset, comprising 5732 labeled images, demonstrate outstanding performance with mAP of 95.7%, while maintaining high detection speed and reliability.