基于灰度相关性的裂缝检测算法

从裂缝的灰度特征入手, 提出一种基于灰度值相似度的裂缝提取方法。针对混凝土路面孔洞、污损等因素干扰, 主要分三步实现裂缝提取：a)对所采集的路面图像进行特征分析, 建立灰度值相似度函数, 根据相似度对比提取疑似裂缝信息; b)利用连通域度量去除虚假裂缝, 将问题转换为图论问题; c）利用谱聚类算法提取真实裂缝。通过大量混凝土路面图片测试表明, 该方法能实现路面真实裂缝特征提取, 具有较强实用性。

Crack detection algorithm based on gray correlation

From the characteristics of the cracks gray, this paper proposed a method of extraction cracks based on gray value similarity to achieve the accurate identification of cracks from the image. For these disturbed factors of pavement holes and defacements, cracks extraction were performed by the following three steps:a)it analyzed the cracks' characteristics of the acquired pavement images, built a similarity function of the gray value and extracted suspected information of the cracks according to the similarity comparison; b)it removed the false cracks by the computed measurement of connected domain and transformed the crack detection problem into graph theory one; c)it extracted the true cracks by the spectral clustering algorithm. By testing many images of concrete pavement, the testing results show that the proposed method can extract the pavement cracks' characteristics actually and have good practicability.