基于Matlab的土体SEM图像处理方法  

Microstructure of soil using SEM images based on Matlab

作　　者：苗得雨[1] 白晓红[1] 

机构地区：[1]太原理工大学建筑与土木工程学院,山西太原030024

出　　处：《水文地质工程地质》2014年第6期141-146,共6页Hydrogeology & Engineering Geology

基　　金：国家自然科学基金资助项目(51178287);山西省自然科学基金资助项目(2010011029);太原理工大学研究生创新基金项目(2013B015)

摘　　要：土的微观结构决定了其物理力学性状,对于土体的微结构研究主要通过对土体的显微图像的定性和定量分析完成。而由于影响土体显微图像成像的因素较多,其往往不能直接用于定量分析,所以需要通过图像处理技术来相对准确地定量表达土体的孔隙和颗粒实际状况。本文利用Matlab,针对土体显微图像的特征,系统提出了Matlab条件下适用于土体显微图像全流程三阶段处理方法:图像预处理、图像分割和优化成像、图像定量统计。通过对比分析处理效果,提出图像不均匀背景消除、对比度调整、图像降噪、图像二值、优化成像、定量分析等方面处理步骤的适用函数和方法,为科研人员进行土体微结构图像定性和定量分析提供技术支持。The physical and mechanical properties of soil are determined by soil microstructure.Soil microstructure study is conducted mainly through the qualitative and quantitative analyses of the SEM images of soil.The SEM images cannot be directly used for quantitative analysis because too many factors may affect its quality.A better quantitative expression of soil pore and grain is obtained through image processing technique and optimization of imaging.By using Matlab,the whole process of image processing is presented in this paper,considering the characteristics of the soil SEM images.The results show a three-segmentation processing：image preprocessing,image segmentation,and optimizing and image quantitative statistics.This paper presents the functions and methods that are appropriate for non-uniform background elimination,contrast adjustment,image noise reduction,image binary,optimize imaging,and quantitative analysis,etc.They can be used to support the qualitative and quantitative analyses of the soil microstructure image.

关 键 词：土体 微结构 MATLAB SEM图像 阈值 

分 类 号：TU411.92[建筑科学—岩土工程]