边缘检测在纸病图像分析中的应用  

The Application of Edge Detection in Paper Defect Image

作　　者：张学兰[1,2] 李军[1,2] 辛培红[3] 李瑞瑞[1,2] 

机构地区：[1]华南理工大学制浆造纸工程国家重点实验室,广东广州510640 [2]华南理工大学广东省造纸技术与装备公共实验室,广东广州510640 [3]华南理工大学电子与信息学院新型人机交互实验室,广东广州510640

出　　处：《造纸科学与技术》2012年第6期121-125,共5页Paper Science & Technology

基　　金：国家重点基础研究发展规划(973计划;2010CB732205);国家科技支撑计划项目(2007BAF25B00)

摘　　要：边缘检测是纸病图像处理的重要方法,也是各类纸病识别的前提。选取孔洞、脏点和褶皱三种有代表性的纸病为研究对象,通过C++编程探究了Roberts算子、Sobel算子和Prewitt算子在纸病图像检测中的效果。通过研究发现,Prewitt算子可以很好地检测出同时含有孔洞、脏点和褶皱的纸病图像的边缘,并且识别率在93.8%以上。为纸病识别的进一步研究奠定了基础。Edge detection is an important method of extracting paper default features, and also the precondition of various kinds of paper defects recognition. This study applied holes, spots and folds to illustrate the effect of Ro- berts, Sobel and Prewitt operator in detecting paper default features based on C ＋ ＋ Experimental results show that Prewitt operator could successfully detect the edge of a paper image that contains more than one kind of paper de- faults such as holes, spots and folds. The performance of the system to detect paper defaults has been demonstrated with 93.8% precision rate. This work is very essential for future studies.

关 键 词：边缘检测 图像处理 纸病图像 

分 类 号：TS77[轻工技术与工程—制浆造纸工程]