一种基于图像融合的岩石裂隙充填物分割方法

岩体的变形特性主要取决于结构面、裂隙，尤其是微裂隙及充填物等方面的共同作用，因此在研究岩石节理时对充填物的研究也非常重要。研究充填物首先须把其图像分割出来。由于彩色图像的噪声太多，以往对岩石裂隙及充填物的分割方法均是基于灰度图像的处理．这不利于对充填物的进一步分析。在处理一个欧盟项目所提供的两种光照图像（普通日光和紫外线光）时，我们发现了一种新的获取岩石裂隙和充填物的方法。该方法基于RGB空间，利用普通日光和紫外线（UV）两种光照系统对岩石样本进行拍照，对所获得的图像分别处理后，再融合、屏蔽掉无关数据，最后利用区域生长法自动在彩色图像上实现了对裂隙充填物的提取。试验证明这是一种高速低噪声的彩色图像分割方法。

A way of segment rock fracture stuffing base on two illuminations images fusion

The characteristics of rock distortion mainly rely on different aspects of rock fractures,such as fracture structure,aperture,and especially tiny apertures and stuffing.The usual method for image segmentation of fracture apertures and stuffing is mainly according to the information of gray level images because of the much noise of color images and it lacks the filling material(stuffing) analysis.The authors have found a novel method for attaining the stuffing and fractures when processing two illumination systems(visible light and ultraviolet waves light) images which were provided by an EU project. The obtained two different color images on the same rock samples.Before image fracture delineation,the noise on the images was removed by the fusion of the information of the two images.Finally,the fractures and stuffing were automatically extracted by using the algorithm based on region growing and edge tracing.As is tested,the method is satisfactory.