Road Marking Line Recognition and Detection System Based on SDF-YOLOv8 in Foggy Weather

The detection and recognition of road markings play a crucial role in road safety maintenance, traffic flow improvement, and the development of autonomous driving technologies. Traditional road marking detection methods based on edge detection and threshold segmentation techniques in image processing generally perform well under clear image conditions. However, in such a situation as the thick fog and other adverse conditions, the degradation of image quality, which leads to reduced contrast and increased noise interference, significantly diminishes their detection performance. In order to solve the above technical challenges, this study constructed an enhanced SDF-YOLOv8 (SCConv-Dyhead-FocaModulationYOLOv8) model based on the efficient real-time detection framework of YOLOv8. By integrating a multi-scale feature pyramid module and a dual-path attention mechanism, the model effectively enhances its feature representation capability for blurred markings and its robustness against complex backgrounds. The experimental results demonstrate that the SDF-Yolov8 model exhibits significant superiority over other models in key performance metrics. While retaining the lightweight advantage of YOLOv8n, it achieves a high recall rate, with precision improving by $2.0 \%$, mAP50 increasing by $3.4 \%$, and mAP50-95 increasing by $7.8 \%$. This research can significantly improve traffic safety and traffic efficiency under foggy weather conditions, providing more accurate environmental perception capabilities for autonomous driving systems.