BGSL_YOLOv8: object detection algorithms for road damage

The detection of road damage is crucial to the development of the national economy and the construction of people's well-being, which holds considerable practical significance and application value. This paper proposes an improved road damage detection algorithm based on YOLOv8, which aims to provide an accurate and stable method for road damage detection. Firstly, this paper newly designs and constructs the BOT_GELAN module, to enable the model to capture information at both local and global levels with greater precision, and improve the accuracy of model detection. Secondly, this paper improves the SPPF module by using LSKA, so that the two work together to enhance the model's capacity to capture local dependence and establish remote dependence, and has both spatial adaptivity and channel adaptivity, so as to enhance the model's detection accuracy. In the experimental results, the proposed model outperforms the original model with a 3.2% improvement in F1-score and a 3.9% increase in mAP.