基于数字图像处理的矿物颗粒形态定量分析  

QUANTITATIVE METHOD OF SHAPE PARAMETERS OF MINERAL PARTICLES BASED ON IMAGE PROCESSING

作　　者：陈建湟 张中俭[1] 徐文杰[2] 李丽慧[3] CHEN Jianhuang;ZHANG Zhongjian;XU Wenjie;LI Lihui(School of Engineering and Technology,China University of Geosciences,Beijing 100083,China;State Key Laboratory of Hydroscience and Engineering,Tsinghua University,Beijing 100084,China;Key Laboratory of Shale Gas and Geoengineering,Institute of Geology and Geophysics,Chinese Academy of Sciences,Beijing 100029,China)

机构地区：[1]中国地质大学(北京)工程技术学院,北京100083 [2]清华大学水沙科学与水利水电工程国家重点实验室,北京100084 [3]中国科学院地质与地球物理研究所,中国科学院页岩气与地质工程重点实验室,北京100029

出　　处：《工程地质学报》2021年第1期59-68,共10页Journal of Engineering Geology

基　　金：国家重点研发计划(资助号:2019YFC1520600);国家自然科学基金(资助号:41602329)。

摘　　要：矿物颗粒形态的评价一般采用目估法在显微镜下观测岩石薄片,其评价结果受主观影响。本文利用岩石薄片显微照片,基于图像处理技术提出了矿物颗粒形态定量分析方法。矿物颗粒形态参数主要包括球度、凸度、长宽比、磨圆度等。该方法利用图像分割技术将显微照片中的矿物颗粒区分开,提取颗粒的像素轮廓坐标进行离散几何分析,计算矿物颗粒形态参数。特别地,由于磨圆度难以直接、准确地计算,本文总结前人经验,在傅里叶级数拟合颗粒轮廓的基础上,创新性地提出了棱角关键点识别和分组两步走的方法。本文分别计算了一组理论图形和一张岩石薄片显微照片的颗粒形态参数,验证了该方法的准确性和实用性。本文分析了图像分辨率Re、颗粒轮廓的傅里叶级数拟合优度R2、棱角关键点分组系数a对颗粒形态参数计算结果的影响,发现:(1)球度、凸度、长宽比的计算结果基本不受Re和R2的影响;(2)磨圆度的计算结果受Re和R2的影响,建议颗粒最小外接圆直径像素数大于200 pixel,并且存在一个最优的R2范围,在这个范围内计算的磨圆度结果接近理论值;(3)当a精确到十分位时,初步分组时选取a=0.1最为合理。The shape characteristics of mineral particles are conventionally evaluated by thin section observation and manual identification of rock samples,the results are often subject to subjective influence.The paper presented a method for quantifying the shape parameters of mineral particles in photomicrographs of rock slices based on an image processing technology.The shape parameters of mineral particles mainly includes sphericity,convexity,aspect ratio,and roundness.Firstly,the method distinguishes mineral particles in photomicrographs using image segmentation technology,then extracts the pixel contour coordinates of the particles for discrete geometric analysis,and finally calculates the shape parameters of mineral particles.In particular,since roundness was difficult to be calculate directly and accurately,the paper proposed a two-step method to calculate the roundness based on Fourier series fitting particle contour,which includes identification the key points in corners and grouping them.The identification of the key points includes two process,which are smoothing particle contour based on Fourier series and searching for the key points in corners.The grouping of the key points also includes two process.Firstly,the key points were preliminarily divided into several groups by statistical analysis,then,the dichotomy was adopted for the precise grouping of these key points.We analyzed the effects of image resolution Re,goodness of fitting Fourier series R2 of particle profile and grouping coefficient a of key points on the calculation results of particle shape parameters.We found that(1)the results of sphericity,convexity and aspect ratio are basically not affected by Re and R2;(2)the results of roundness are affected by Reand R2.It is suggested that the minimum circumscribed circle diameter of particles should be more than 200 pixels for roundness calculation,and there is an optimal R2 range,within which the calculated roundness is close to the theoretical value;(3)when a is accurate to the tenth,a=0.1 is the most

关 键 词：岩石薄片 图像处理 矿物颗粒 形态量化 形态参数 傅里叶级数 

分 类 号：P642.3[天文地球—工程地质学]