Foggy Road Scene Detection Based on Improved Yolov8n

In response to the target detection of intelligent systems under foggy conditions, which faces the problem of reduced detection accuracy due to reduced contrast of captured images, color shift and loss of details, a foggy road scene detection method based on improved Yolov8n is investigated. Firstly, the DehazeNet algorithm is combined with the target detection algorithm for defogging to obtain relatively clear images in order to reduce the interference of fog on target detection; Secondly, the Dilation-wise Residual (DWR) module is added to a portion of the C2f module of the backbone network, which enables the network to flexibly adapt to features at different scales to improve the detection accuracy; Finally, the proposed method is tested on the real fog-containing road scene dataset RTTS, and the mAP@50 is improved by 2 percentage points compared with the original model, which validates the feasibility and effectiveness of the design ideas and the proposed method in this paper.