基于相位一致性图像处理的岩体裂隙自动检测

Automatic detection of rock mass fissure based on image processing of phase congruency

提出了一种基于相位一致性检测和分段线迭代连接并去除伪边缘的岩体裂隙自动检测方法。先对原始图像进行相位一致性边缘检测并细化处理,得到包含裂隙特征和伪边缘的分段线,再根据距离和角度原则对分段线迭代连接,裂隙分段线相互连接而不断完整的同时滤除不能连接而变长的伪边缘。实验表明,对于大体积岩体露头面图像的裂隙检测,该方法具有较高的精确性。

A new method is presented for detection of rock mass fissure, which is based on phase congruency and iterative connection of segmented lines with removing of pseudo edge. Firstly, detecting the image by phase congruency and thinning the outcome, so the segmented lines including the crack feature and the pseudo edge are obtained. And then, connecting the segmented lines iteratively according to the distance and the angle principle, making the segmented lines be continuously, and filtering the pseudo edges which can not lengthen. Experiments show that for the detection of a large volume of rock mass fissure, this method is more accurate than others.