基于改进U-Net网络的路面裂缝检测方法

打开文本图片集

关键词：路面裂缝检测；U-Net;轻量化网络；注意力机制;多尺度特征融合;空洞卷积;特征金字塔网络；智能养护

中图分类号：TP391.4文献标志码：Adoi：10.12415/j.issn.1671-7872.25069

Abstract： A lightweight network named MSAL-UNet， based on an improved U-Net， was proposed to address the challenge of balancing accuracyand computational efficiency in pavement crack detection.First， a lightweight encoder wasconstructedand integrated with the convolutional block atention module（CBAM） to dynamically enhance feature responses in key crack regions，thereby improving the model's perception capability for low-contrast targets.Second，a multi-scale dilated convolution （MSDC） module was designed to expand the receptive field without reducing spatial resolution， efectively capturing multi-scale information from fine cracks to broad contextual features and significantly suppressing false positives.Furthermore，a feature pyramid network （FPN） was introduced to achieve cros-level semantic alignment and multi-scale feature fusion，enhancing the coherence of crack localization and boundary smoothness. Comparative and ablation experiments were conducted on the public dataset CRACK500 to validate the performance of the proposed method.The results demonstrate that the proposed method achieves a mean intersection over union （mIoU）of 0.832 0 and a precision of 0.918 0（the highest among all compared models），with anF1-score of 0.9063， while maintaining a real-time inference speed of 34.45 frames per second.The comprehensive performance is significantly superior to mainstream lightweight networks such as BiSeNet，Fast-SCNN，SegFormer，andTransUNet.Particularlyinscenarios involving faintandcomplexcracks，the segmentation results of MSAL-Unet exhibit optimal performance in both completeness and boundary accuracy. By synergistically optimizing multi-scale feature fusion，attention mechanisms，and lightweight structural design， this study effectively addresses the challenging balance between accuracy and efficiency in pavement crack detection， thereby providing highlyreliable technical support with low computational overhead for practical road intellient inspection systems.

Keywords： pavement crack detection; U-Net; lightweight network; attention mechanism; multi-scale feature fusion： dilated convolution; feature pyramid network; intelligent maintenance

随着我国社会经济持续发展和新型城镇化深入推进，公路交通出行需求迅猛增长。（剩余15189字）