摘要
为了探讨K-means算法应用于图像分割时在不同颜色空间中的聚类效果,选用了不同分辨率的多对图像进行研究,分析了基于RGB和YUV颜色空间的分割结果,并提出一种新的混合模型,即在YUV聚类距离公式中引入图像的二维信息熵的差量,同时利用YUV颜色空间中的Y分量作为其灰度进行计算,实验结果表明,基于YUV颜色空间聚类的改进模型分割效果比单纯使用YUV颜色空间进行聚类更佳。
In order to investigate the clustering effect of K-means in different color space when applied in image segmentation, a series of studies were carried out and pairs of images with different resolution were used to do the test. The effects of clustering in RGB ＆ YUV color space were analyzed and a new combination model was proposed by introducing two dimensional information entropy differential into YUV clustering distance equation and utilizing Y component to make gray scale calculation. The result shows that new improved model achieved the better seg- mentation effect than only using clustering based on YUV color space.
出处
《太原理工大学学报》
CAS
北大核心
2014年第3期372-375,共4页
Journal of Taiyuan University of Technology
基金
山西省科技攻关项目(20130313030-1)
关键词
图像分割
RGB颜色空间
YUV颜色空间
K-均值聚类
二维信息熵
Image segmentation
RGB color space
YUV color space
K-means value clustering
Two dimensional information entropy
作者简介
作者简介：王爱莲（1975-），女，山西孝义人，讲师，博士，主要从事SNS数据分析，数据通信，无线传感器研发及无线数据分析的研究，（E—mail）jb636766@sina．com
