基于最小代价路径搜索的路面裂缝检测

Pavement Crack Detection Based on Minimum Cost Path Searching

针对路面影像斑点噪声强,路面裂缝对比度低、连续性差等特点,提出了一种基于最小代价路径搜索的路面裂缝检测方法。在裂缝成像亮度低于路面背景的前提下,该方法将裂缝提取问题转化为格状图顶点之间最小代价路径搜索的问题,首先通过对裂缝面元的处理,获取潜在的裂缝种子点,然后运用基于最小代价路径搜索的算法,实现裂缝种子点的生长和连接,并最终提取裂缝。试验结果表明:所提方法具有较高的精度和效率。

Aimed at character of high speckle noises of pavement images,low contrast and bad continuity of pavement cracks,pavement crack detection approach based on minimum cost path searching was proposed.In this approach,the crack detection problem was formulated as a minimum cost path searching problem under assumption that cracks were imaged with a lower intensity than pavement background.Firstly,potential crack seed points were collected by processing of the crack block.Then,the crack seed points were linked into a whole crack through an algorithm on the basis of minimum cost path searching algorithm.Experiment results show that the proposed method has high accuracy and efficiency.