基于三维光影模型的公路路面裂缝自动识别算法  

Automatic Detection Method for Highway Pavement Cracking Based on the 3D Shadow Modeling

作　　者：阳恩慧[1,2] 张傲南[1,2] 丁世海[1,2] 王郴平[1,2] 

机构地区：[1]西南交通大学土木工程学院,四川成都610031 [2]道路工程四川省重点实验室,四川成都610031

出　　处：《西南交通大学学报》2017年第2期288-294,共7页Journal of Southwest Jiaotong University

基　　金：国家自然科学基金资助项目(51308477;U1534203);中央高校基本科研业务费专项资金资助项目(2682015CX091)

摘　　要：针对公路路面裂缝的自动化三维图像识别技术的研究热点问题,为有效提高裂缝识别算法的准确率与可靠性,基于三维图像提出利用三维光影模型(3D Shadow Modeling)来实现公路路面表面裂缝的自动识别新方法.该方法利用裂缝区高度低于周边高度的特性,通过三维光影模型将裂缝投影为阴影区,继而通过对阴影的形态分析来识别裂缝,并采用连通域分析与线性形态分析方法,消除图像噪声.研究结果表明,采用本文提出的裂缝自动识别算法可以达到92.93%的路面裂缝自动识别准确率.Automatic detection method for pavement cracking based on 3D image technology has been a hot topic. In order to improve the accuracy and reliability of the cracking detection, a new method based on the 3D shadow modeling was proposed, which utilizes that the height of the fracture zone is lower than the surrounding area, and the shadow model was used to identify the cracking by the virtual light projection and then, the linear shape analysis method was used to discriminate cracks detection. Furthermore, both the connected-component analysis method and linear shape analysis method were adopted to remove the image noises. The results show that the automatic detection method can achieve 92.93% of accuracy in pavement cracking detection.

关 键 词：三维光影模型 公路路面 裂缝 识别算法 

分 类 号：U416.22[交通运输工程—道路与铁道工程]