改进单尺度Retinex算法在图像增强中的应用  

APPLICATION OF IMPROVED SINGLE SCALE RETINEX ALGORITHM IN IMAGE ENHANCEMENT

作　　者：陈超[1] 

机构地区：[1]四川理工学院计算机学院,四川自贡643000

出　　处：《计算机应用与软件》2013年第4期55-57,74,共4页Computer Applications and Software

基　　金：国家自然科学基金项目(51108390);人工智能四川省重点实验室开放基金项目(2011RYY06);四川理工学院国家培育项目(2011PY05)

摘　　要：为了提高光照条件变化下的图像增强效果,提出一种改进单尺度Retinex(SSR)算法的图像增强方法。利用双边滤波代替SSR算法的高斯滤波对光照估计,克服SSR算法不能解决的"光晕"问题。结果表明,相对传统图像增强算法,改进SSR算法不仅加快了图像增强处理的效率,且获得了更好的图像增强处理效果,是一种适应性较强的、可用于光照变化比较强烈下的图像增强算法。In order to improve image enhancement efficiency in illumination variation,this paper proposes an image enhancement with improved single scale Retinex algorithm（SSR）.Bilateral filtering is used instead of Gauss filtering of SSR algorithm in illumination estimation to overcome the problem of halo artifact which SSR algorithm can not solve.Results show that the improved SSR algorithm can accelerate the image enhancement processing efficiency compared with conventional image enhancement algorithms,and also it obtains a better processing effect,and is an image enhancement algorithm with stronger adaptivity and can be used in stronger illumination variation situation.

关 键 词：图像增强 RETINEX算法 双边滤波 光晕现象 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]