纸病图像的数学形态学分割方法  

THE SEGMENTATION OF SURFACE DEFECTS IMAGE BASED ON MATHEMATICAL MORPHOLOGY

作　　者：吕岑[1] 张根宝[1] 孙瑜[1] 

机构地区：[1]陕西科技大学电气与电子工程学院,陕西咸阳712081

出　　处：《陕西科技大学学报（自然科学版）》2003年第6期82-85,共4页Journal of Shaanxi University of Science & Technology

基　　金：国家外专局引智专项计划项目(WJ2002067);陕西省教育厅产业化培育项目(02JC17);陕西科技大学自然科学专项(ZX02-41)资助项目。

摘　　要：提出了采用数学形态学操作从纸病图像中分割强噪声背景和纸病的方法。图像噪声和纸病的两个明显的区别是它们的密度和尺寸大小,利用噪声和纸病这些明显区别的特性,采用形态学的腐蚀、膨胀、开运算及闭运算等操作分两次去除图像噪声、检测图像边缘,结合阈值增强方法和传统边缘检测得到纸病检测和分割结果。The method of segmenting the background noise and surface defects from the paper images by the mathematical morphology is proposed. Two distinct differences between image noise and defects are density and scale. By these characteristics, using morphology′s operators, such as erosion, dilation, opening and closing operation to remove the image noise two times and detect the defects edges. Experimental results of defects segmentation and detection obtained by applying the traditional thresholding and edge detector are presented.

关 键 词：数学形态学 图像处理 纸病检测 边缘检测 图像分割 

分 类 号：TS77[轻工技术与工程—制浆造纸工程] TS736.2