岩石节理裂隙骨架提取及其整体分形分析

Rock fracture skeleton extraction and fractal analysis

在对彩色岩石节理裂隙图像或灰度图像进行分割后,二值图像的骨架结果一般很难让人满意,因为图像中的各种噪声和线段的不连续性使测量困难并影响测量的精度.根据曲线的细线化、端点和节点的检测等设计并实现了一种基于二值图像形态学特性的曲线去噪、断线连接及整体曲线分形拟合的方法,改善了以往检测出的节理裂隙曲线具有不连续性、带有毛刺噪声及整体复杂度的判断问题,从而可以对每条裂隙曲线进行分形测量,不仅仅进行了单一曲线的分形描述,还进行了多曲线组合的分形描述,分形结果可作为评定整体节理裂隙的特征之一.

After color or grey level rock fracture image processing and segmentation, the binary image result is not satisfactory in most cases, due to the different image noise and gaps of curves or lines, which makes fracture measurement difficult and affects the measurement accuracy. This paper designed and realized a method for tracing rock fractures and removing image noise according to line or curve thinning, end point and junction point detection, and B - Spline techniques. The studied method improved the rock fracture measurement accuracy since the noise in the image and on the fractures were removal, the gaps on lines or curves were sewing up, and the rugged or rough lines or curves were smoothed by using B -Spline. In this way, the fracture measurement accuracy can be increased by using fractal analysis, and in this paper the fractal analysis is not only for singlke curve, and also for multiple curves in one image.