摘要
沥青路面裂缝作为道路典型病害之一,裂缝的自动识别技术在道路检测行业需求极为迫切,但现有的识别算法难以达到工程应用级精度,其主要原因之一是由于路面裂缝图像未有完备的标注规则及数据处理机制。本研究通过对裂缝病害判断规则的制定,到路面裂缝的判断依据,据此构建裂缝标准图像训练集,通过完备的数据图像处理流程,可实现路面裂缝较高的识别准确度,为扩大路面裂缝自动识别项目级应用奠定基础。
As one of the typical road diseases,the automatic identification technology of asphalt pavement is extremely urgent in the road detection industry,but the existing identification algorithm is unable to achieve engineering application level accuracy,one of the main reasonsis that the pavement crack image does not have complete labeling rules and data processing mechanism.Based on the formulation of crack disease judgment rules and the judgment basis of road cracks,this study constructed a training set of standard crack images.Through the complete data image processing process,a relatively high identification accuracy of road cracks can be achieved,thus laying a foundation for expanding the project-level application of automatic identification of road cracks.
作者
韩海航
张丽
吴皓天
陈海珠
胡迪
HAN Haihang;ZHANG Li;WU Haotian;CHEN Haizhu;HU Di(不详)
出处
《科技创新与应用》
2021年第9期29-34,共6页
Technology Innovation and Application
基金
浙江省交通运输厅科技计划项目(编号:2020053)资助。
关键词
图像处理
沥青路面
裂缝
病害识别
image processing
asphalt pavement
crack
disease recognition
