沥青混合料CT图像分割技术及其效果分析  

Image Segmentation and Its Effect of Asphalt Mixtures Using Computed Tomography Images Method

作　　者：刘佳辉[1] 李智[1] 

机构地区：[1]华南理工大学土木与交通学院,广东广州510640

出　　处：《重庆交通大学学报（自然科学版）》2011年第6期1335-1338,共4页Journal of Chongqing Jiaotong University(Natural Science)

基　　金：湖南省交通厅科技进步与创新计划项目(200809);广东省交通厅重点工程项目(2009-01-003);华南理工大学中央高校基本科研业务费资助项目(2009ZM0099)

摘　　要：在最大类间方差法(OTSU法)的基础上,设计了环状分块方法,用以解决在处理路面芯样CT图像灰度图无明显双峰,目标集料颗粒与沥青胶浆背景分割的技术难题,并着重从图像分割效果方面设计试验,开展了评价研究。结果表明:采用环状分块和OTSU法相结合的改进算法,图像的分割质量明显提高。进一步与最佳阈值法(手动)相比较,采用粗集料颗粒面积比作为指标,定量地评价了环形分块OTSU法的精度,结果理想。A circular blocking method is designed on the basis of OTSU method.This method is proposed to resolve the technical problems of segmentation between target aggregate particles and asphalt mortar background in CT image gray of pavement core samples without obvious bimodal.Moreover,evaluation studies are carried out through designing experiments emphatically from image segmentation.The results show that the quality of image segmentation has been improved significantly by using circular blocking method and OTSU method together.Furthermore,comparing with the optimal threshold method（Manual）,the accuracy of the circular block OTSU method is evaluated quantitatively by using area ratio of aggregate particles as an indicator,and the result turns out to be ideal.

关 键 词：沥青混合料 图像分割 环形分块 OTSU法 

分 类 号：U414.1[交通运输工程—道路与铁道工程]