基于分数阶微分和聚类分析的矿岩三维图像裂缝提取算法

Extraction Algorithm for Fractures in Mineral Rock 3D Images Based on Fractional Order Differential and Cluster Analysis

针对矿岩三维图像存在较大阴影和非均匀背景导致浅层裂缝容易误判的问题,提出了一种基于分数阶微分和聚类分析的矿岩三维图像裂缝提取算法。首先采用分数阶微分最小邻域值方法,消除图像背景噪声并增强裂缝边缘信息;其次改进MASK匀光算法,提升算法的自适应能力,进一步用于提升图像对比度;最后采用K-means聚类分析算法对矿岩三维图像进行裂缝提取。试验结果表明:该方法能够有效降低低光照情况下阴影对裂缝提取的影响,对噪声具有平滑效果,实际矿岩三维图像裂缝提取的准确率达到了91.5%,为矿岩图像信息提取与判读分析提供了有效方法。

Aiming at the problem that the shallow crack is easy to misjudge due to the large shadow and non-uniform background in the 3D image of mineral rock,a crack extraction algorithm based on fractional differentiation and cluster analysis is proposed in this paper.Firstly,the fractional differential minimum domain value method is used to eliminate the background noise and enhance the crack edge information.Secondly,the MASK smoothing algorithm is improved to improve the adaptive ability of the algorithm and further improve the image contrast.Finally,K-means clustering analysis algorithm is adepted to extract cracks from 3D images of mineral rocks.The experimental results show that this method can effectively reduce the influence of shadow on crack extraction under low light conditions,and has a smooth effect on noise.The accuracy rate of crack extraction in actual 3D images of mineral rock is up to 91.5%,which provides an effective method for image information extraction and interpretation analysis of mineral rock.