有效消除光晕现象和颜色保持的彩色图像增强算法  

A Halo-Free and Hue Preserving Algorithm for Color Image Enhancement

作　　者：肖泉[1] 丁兴号[1,2] 王守觉[1,3] 廖英豪[1] 郭东辉[1] 

机构地区：[1]厦门大学信息科学与技术学院,厦门361005 [2]厦门大学水声通信与海洋信息技术教育部重点实验室,厦门361005 [3]中国科学院半导体研究所神经网络研究室,北京100083

出　　处：《计算机辅助设计与图形学学报》2010年第8期1246-1252,共7页Journal of Computer-Aided Design & Computer Graphics

基　　金：国家自然科学基金(30900328);福建省自然科学基金(2008J0032;2009J01301;2009J01302);厦门大学985二期信息创新平台资助项目(0000-X07204);厦门市科技计划高校创新项目(3502Z20083006)

摘　　要：由于对图像中明暗突变区域的背景光照估计不准确,经典Retinex彩色图像增强算法易产生光晕现象且存在增强后图像细节信息减弱和颜色失真等不足.为此,结合人眼视觉特性提出一种彩色图像增强算法.首先利用人眼对图像结构特征及颜色信息的敏感特性,通过构造彩色双边滤波器来获取图像背景光照,以避免光照突变处产生光晕现象;其次依据人眼系统局部自适应调节特性,通过引入一个对比度调节函数自适应增强图像的细节信息,克服经典Retinex算法在整体对比度提高的同时局部对比度下降的不足;最后利用一种线性的颜色恢复算法恢复增强所得亮度图像的颜色信息.与MSRCR等彩色图像增强算法比较的实验结果表明,文中算法更有效,增强后的图像不仅细节清晰,而且色彩鲜艳、自然.As the estimation of background luminance at the mutation of light is usually inaccurate, the traditional Retinex enhancement algorithms suffer from the halo phenomenon, as well as the loss of details and color distortion in the enhanced image. Based on the characteristics of the human visual system, a novel color image enhancement algorithm is proposed in this work. As the human visual system is sensitive to structural features and color information of image, a color bilateral filter is constructed to estimate the background luminance, which can effectively overcome the halo phenomenon. Moreover, by using the local self-adjustment characteristic of the human visual system, a local contrast enhancement function is introduced to adaptively adjust the intensity of each pixel, to overcome the dilemma that the overall contrast is improved but the local contract is reduced. Finally, a color restoration process is utilized to convert the enhanced intensity image back to the color image. Experimental results show that the proposed algorithm is more effective in terms of visual effects, and the enhanced image is not only more detail-preserving, but it is also more colorful and natural, compared with other methods such as MSRCR.

关 键 词：图像增强 人眼视觉系统 RETINEX 彩色双边滤波 局部对比度 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]