一种基于粗糙集合理论的彩色图像分割  

Color Image Segmentation Based on Rough Set Theory

作　　者：张恒[1] 冯子亮[1] 

机构地区：[1]四川大学图像图形研究所,四川成都610064

出　　处：《计算机技术与发展》2009年第2期39-41,44,共4页Computer Technology and Development

基　　金：国家自然科学基金重点资助项目(60736046)

摘　　要：在数字图像处理中,图像分割是其中的一个重要问题,也是一个经典难题。基于阈值选取方法的图像分割计算简单,具有较高的运算效率,其方法多种多样,但是方法选取直接影响到图像分割的质量。为了对彩色图像进行合理的分割,获得合适的阈值,提出了一种基于粗糙集理论的彩色图像分割算法。它将基本直方图作为图像粗糙集的下近似,将Histon直方图作为其上近似,计算出图像粗糙度后通过单峰子集分离法实现对图像的分割。实验结果表明该算法对背景复杂和颜色渐变的彩色图像有较好的分割效果。Image segmentation is an important problem of digital image processing and also a common difficult problem. Image segmentation based on selection of image threshold methods has characters of calculating simply and operating efficiency highly. The efficiency of image segmentation depends on the selection of image threshold methods directly. For segmenting natural images with regions having gradual variations in color value, a new technique based on the rough set theory for color image segmentation has been presented in this paper. By considering the histogram as the lower approximation of the rough set of image and the Histon as upper one, the segmentation of color images can be processed by the method for separation of using unimodal distribution. The experimental results show that the technique has excellent effect on the images with complex background or gradually changing shades.

关 键 词：粗糙集 直方图 图像分割 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]