阈值分割法处理土微结构图像质量评价  

Quality Evaluation of Soil Micro-structure Images Processed by Thresholding

作　　者：方磊[1] 李春苗[1] 龚文宗[2] 王军培 

机构地区：[1]东南大学交通学院,南京210096 [2]中交四航工程研究院有限公司,广州510230 [3]无锡市建筑设计研究院有限责任公司,江苏无锡2140001

出　　处：《地下空间与工程学报》2013年第S2期1873-1876,共4页Chinese Journal of Underground Space and Engineering

基　　金：国家自然科学基金项目(41172239)

摘　　要：在对常用图像分割质量评价方法及度量参数进行整理和研究的基础上,将均值、标准差、信息熵、空间频率、平均梯度等定量评价参数应用到土微结构图像阈值法分割质量的评价中来,对双峰阈值法、Otsu法、迭代法以及最大信息熵法分割图像的效果进行了评价。实验结果表明,针对本文图像,双峰阈值法分割效果较为理想。On the basis of collecting and researching the commonly used evaluation methods and metrics of Image segmentation quality,the quantitative evaluation parameters such as the mean,standard deviation,information entropy,spatial frequency and the average gradient are applied to evaluate the quality of the soil microstructure images segmented by thresholding. In this way,the effects of image segmentation using bimodal threshold,Otsu,iterative and information entropy methods are evaluated. The results show that the effect of bimodal thresholding segmentation is better for images given in the article.

关 键 词：阈值分割法 土微结构图像 质量评价 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] P642.1[自动化与计算机技术—计算机科学与技术]