彩色影像的遗传自适应增强  

Genetic Adaptive Color Image Enhancement

作　　者：潘励[1] 张祖勋[1] 张剑清[1] 

机构地区：[1]武汉大学遥感信息工程学院,武汉市珞喻路129号430079

出　　处：《武汉大学学报（信息科学版）》2001年第3期253-255,274,共4页Geomatics and Information Science of Wuhan University

基　　金：国家自然科学基金!资助项目 (498710 6 7) ;测绘遥感信息工程国家重点实验室开放研究基金!资助项目 (WKL(0 0 ) 0 10 1)

摘　　要：提出了一种彩色影像自适应增强的算法 ,此算法充分利用了彩色影像饱和度和亮度所包含的信息 ,并利用遗传算法自适应地调整增强系数。对于不同的影像 ,本文算法均能使其对比度、目标边缘以及纹理特征得到增强。Image enhancement is one of the important image processing techniques.It is used to improve image quality or extract the fine details in the degraded images.For color image enhancement,most existing enhancement techniques make use of the luminance,hue and saturation description of a color image.Because the saturation component often contains higher frequency spectral energy,i.e.image detail,than its luminance counterpart.A number of researchers present to feed back high_pass information from the saturation component as a means of supplementing color image sharpness and contrast.This technique of “saturation feedback' can serve to bring out image details that have low luminance contrast.Based on the technique,a lot of enhancement approaches have been proposed.But these approaches are parameter dependent.In other words,for each image,the author has to manually adjust feedback parameters so as to obtain satisfied enhancement images.Obviously,it is too troublesome to satisfy real time image processing. In this study,a genetic algorithm approach to color image enhancement is proposed,in which color image enhancement is formulated as an optimization problem.The genetic algorithm is an adaptive procedure that searches for good solutions by using a collection of search points known as a population in order to maximize some desirable criterion.Also,the genetic algorithm,as a stochastic random search technique,is known to use the accumulating information to prune the search space,while a purely random search ignores information about the environment.In the proposed approach,feedback parameters are optimized components.The fitness function for genetic algorithm is formed by fuzzy entropy and fuzzy contrast.Then genetic algorithm is used to determine the “optimal' feedback parameters with the largest fitness function value.This paper also discusses the detail procedures of encoding,selection,crossover and mutation in the genetic algorithm. Experiments are done on color aerial images.Based on the experimental result

关 键 词：彩色影像 饱和度 亮度 遗传算法 对比度 目标边缘 纹理 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置] TP183[自动化与计算机技术—控制科学与工程]