An Improved YOLOv5 Crack Detection Method Combined with a Bottleneck Transformer

Efficient detection of pavement cracks can effectively prevent traffic accidents and reduce road maintenance costs. In this paper, an improved YOLOv5 network combined with a Bottleneck Transformer is proposed for crack detection, called YOLOv5-CBoT. By combining the CNN Transformer, YOLOv5-CBoT better capture long-range dependencies to obtain more global information, so as adapt long-span task cracks. Moreover, C2f module, which in state-of-the-art object YOLOv8, introduced further optimize by paralleling gradient flow branches richer information. The experimental results show that has achieved competitive on RDD2020 dataset, fewer parameters lower computational complexity but higher accuracy faster inference speed.