Journal of Computer Applications ›› 2012, Vol. 32 ›› Issue (09): 2573-2575.DOI: 10.3724/SP.J.1087.2012.02573

• Graphics and image technology • Previous Articles     Next Articles

Enhancement algorithm for color images based on co-occurrence matrix

YANG Bing-qing1*,TIAN Xiao-ping2,WU Cheng-mao2   

  1. 1.School of Communication and Information Engineering,Xi'an University of Posts and Telecommunications,Xi'an Shaanxi 710061,China;
    2.School of Electronic Engineering,Xi'an University of Posts and Telecommunications,Xi'an Shaanxi 710061,China
  • Received:2012-03-23 Revised:2012-05-11 Online:2012-09-01 Published:2012-09-01

基于共生矩阵的彩色图像增强算法

杨冰清1*,田小平2,吴成茂2   

  1. 1.西安邮电大学 通信与信息工程学院,西安710061;
    2.西安邮电大学 电子工程学院,西安710061
  • 通讯作者: 杨冰清
  • 作者简介:杨冰清(1986-),女,陕西韩城人,硕士研究生,主要研究方向:图像与视频通信; 田小平(1963-),男,陕西渭南人,教授,主要研究方向:图像处理、图像通信; 吴成茂(1968-),男,四川仪陇人,高级工程师,主要研究方向:图像处理。
  • 基金资助:

    陕西省自然科学基金资助项目(2009JM8004);陕西省教育厅科技计划项目(2009JK730,2010JK816)

Abstract: Considering that the traditional method of histogram equalization can make the image produce unnatural enhanced results, a new color image contrast enhancement approach was proposed, which equalized color components by dependently using co-occurrence matrix. Firstly, gray correlative characteristics of pixels were combined in 3×3 neighborhoods to construct a co-occurrence matrix. Secondly, the method of co-occurrence histogram equalization was applied to luminance component only and the chrominance components were preserved. Finally, dark channel prior was used to solve the problem mentioned above to obtain the ideal image. Comparing the proposed enhanced algorithm with the other three typical enhancement algorithms, it is shown that the proposed algorithm not only considers the whole and local image information, but also deals with the histogram spikes and halo effect. The experimental results demonstrate that the proposed method can enhance the color images effectively.

Key words: image enhancement, co-occurrence matrix, dark channel prior, halo effect, definition

摘要: 针对传统的直方图均衡化会使图像过分增强,产生不自然的现象,提出一种利用二维共生矩阵的彩色图像增强算法。首先,利用相邻3×3像素之间的相关性来构造共生矩阵;然后,对彩色图像中的亮度分量进行共生直方图均衡;最后,将增强后的图像进行暗原色处理得到比较理想的图像。从所提增强算法与其他3种典型的增强算法的比较可看出:所提算法不仅考虑了图像的全局与局部信息,而且降低了直方图峰值和光环效应,提高了图像对比度,有较好的增强效果。

关键词: 图像增强, 共生矩阵, 暗原色先验, 光环效应, 清晰度

CLC Number: