基于边缘形态变换的彩色织物图像分割算法  

Color fabric image segmentation based on the edge morphological transform

作　　者：赵海英[1,2] 彭宏[1] 杨一帆[2] 

机构地区：[1]新疆师范大学计算机科学技术学院,乌鲁木齐830054 [2]北京科技大学信息工程学院,北京100083

出　　处：《中国体视学与图像分析》2011年第1期5-10,共6页Chinese Journal of Stereology and Image Analysis

基　　金：国家自然科学基金(60863010);973前期计划专项课题(2010CB334709);新疆自然科学基金(2010211a19)

摘　　要：新疆民间织物图像具有鲜明的民族特色,为了生成和创新民族民间织物图像,如何自动分割织物基元是一个关键问题。论文以问题为驱动,提出了一种基于边缘形态变换的彩色织物图像分割算法。首先,对彩色图像进行颜色空间的变换,然后根据图像特征提出调整图像大小的前处理标准,再利用Canny算子检测边缘,对检测边缘进行形态变换后处理,最后标记变换图像的封闭区域进行织物图像的分割。实验数据表明:该算法虽然是几种常用算法的组合,但较好解决了民间织物彩色图像分割问题,而且对不同类型的织物图像,均可以获得令人满意的分割结果。Xinjiang folk fabric image has distinct nationality characteristics.There is a key problem that in order to generate and innovation how to segment the folk fabric images.In order to solve this problem,this paper puts forward a segment algorithm based on edge morphological transform for color fabric images.Firstly,color space transformation is implemented.Then the image size norms are proposed according to the characteristics of image,and Canny operator and morphological transform are executed to detect edge of image as post-processing.Finally,based on closed mark transform area fabric image segmentation is carried out.Experimental data show that the algorithm is generated by combination of several common algorithms,but better solved the folk fabric color image segmentation problem,and can be used to obtain satisfactory segmentation results for different types of fabric images.

关 键 词：彩色图像分割 形态变换 边缘提取 织物基元 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]