基于改进模糊C均值聚类和区域合并的矿物颗粒分割方法  

Mineral Particle Segmentation Algorithm Based on Improved Fuzzy C -means and Region Merging

作　　者：朱磊 滕奇志[1] 龚剑 ZHU Lei;TENG Qi-zhi;GONG Jian(Institute of Image Information,College of Electronics and Information Engineering,Sichuan University,Chengdu 610065,China;Chengdu Xitu Technology Co.,Ltd.,Chengdu 610024,China)

机构地区：[1]四川大学电子信息学院图像信息研究所,成都610065 [2]成都西图科技有限公司,成都610024

出　　处：《科学技术与工程》2020年第34期14138-14145,共8页Science Technology and Engineering

基　　金：国家自然科学基金(61372174)。

摘　　要：岩矿石薄片图像中矿物颗粒的分割提取是分析鉴定工作的前提,为了进一步提高矿物颗粒提取的准确性,提出一种新的矿物颗粒提取分割方法。该方法以一组岩矿石薄片正交偏光序列图作为输入,对每张正交偏光序列图采用改进的模糊C均值聚类(fuzzy C-means,FCM)算法进行初始的边缘提取,并叠加多张边缘图像,再对得到的叠加边缘图像进行精细化处理并进行目标颗粒筛选,得到初始的分割提取图像。为了抑制颗粒的过分割现象,利用基于区域邻接图(region adjacency graph,RAG)的合并算法进行区域合并,首先依据颗粒的颜色纹理等特征进行初始合并操作,再根据矿物颗粒在不同正交偏光角度下的变化规律进行二次合并操作,从而得到最终的分割提取图像。通过对多组图像进行实验表明,该方法可以获得良好的效果。Segmentation and extraction of mineral particles in rock ore slice images is the premise of analysis and identification.In order to further improve the accuracy of mineral particle extraction,a new method of mineral particle extraction and segmentation was proposed.A set of orthogonal polarized light sequence images of rock ore slice were taken as input in this method.Firstly,an improved fuzzy C-means(FCM)algorithm was used in each orthogonal polarized light sequence image for initial edge extraction.Secondly,multiple edge images were superimposed.Thirdly,the resulting superimposed edge images were refined and target particles were filtered to obtain the initial segmentation image.Finally,in order to suppress the over-segmentation of the particles,the region adjacency graph(RAG)-based merge algorithm was used to perform region merge.An initial merging operation was performed according to the color and texture characteristics of the particles firstly,and then a secondary merging operation was performed according to the change rule of the mineral particles under different orthogonal polarization angles,so as to obtain the final segmentation image.Experiments on multiple sets of images show that the proposed method can obtain good results.

关 键 词：正交偏光序列图像 图像分割 模糊C均值聚类 区域合并 区域邻接图 卡方比较 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]