摘要
为提升沥青路面裂缝的检测精度,文章利用数据增强技术,将路面图像编码为RGB数组并进行预处理,通过迭代聚类计算确定最佳分簇数以增强裂缝特征。实验利用多元数据集,对比使用技术前后数据,准确率、召回率均得到提升,均方误差降低,为裂缝检测提供了可靠支持,提升了道路维护水平。
To enhance the detection accuracy of asphalt pavement cracks,this paper employs data augmentation techniques to encode road images into RGB arrays and preprocess them.The optimal number of clusters is determined through iterative clustering calculations to enhance crack features.Experiments using a diverse dataset show that both accuracy and recall rates are improved,and the mean square error is reduced.This study provides reliable support for crack detection and improves road maintenance levels.
作者
韩永康
HAN Yongkang(Zhouqu Highway Section,Gannan Highway Development Center,Zhouqu 746300,China)
出处
《科技创新与生产力》
2025年第4期133-136,共4页
Sci-tech Innovation and Productivity
关键词
沥青路面
裂缝检测
数据增强
聚类分析
像素下降率
asphalt pavement
crack detection
data augmentation
cluster analysis
pixel decrease rate
