非均匀光照图像的子块部分重叠对数变换  

Partially overlapped sub-block logarithmic transformation of images under non-uniform illumination

作　　者：焦立男[1] 孙朝云[1] 沙爱民[2] 

机构地区：[1]长安大学信息工程学院电子信息工程系,西安710064 [2]长安大学公路学院,西安710064

出　　处：《计算机工程与应用》2011年第9期196-199,共4页Computer Engineering and Applications

基　　金：博士后科学基金项目(No.20100471589);陕西省自然科学基础研究计划项目(No.2009JQ8009);交通部西部交通建设科技项目(No.200831881264)~~

摘　　要：在非均匀光照下获得的图像往往具有高动态范围。大多数现有的全局对比度增强算法对此种图像效果并不好。一种简化的对数变换公式被提出来解决这一问题,更进一步子块部分重叠方法也被采纳进来实现局部的对比度增强。为了达到优化的目的,两种对比度测度用来寻找优化的变换参数。这些方法合并成一种新的算法POSLT(子块部分重叠的对数变换)。POSLT尽管计算复杂性比较高,但是展示出了良好的性能和可控性。实验结果也显示出了该算法在对比度增强中的局部适应性,有力支持了POSLT。Most existing global contrast enhancement algorithms can not deal with high dynamic range images which are often under non-uniform illumination.Thus a simplified logarithmic transformation equation has been proposed to meet this demand.Furthermore a partially overlapped sub-block method is also adopted to perform local contrast enhancement.For purpose of optimization,two contrast measures are used for finding optimal transformation parameters.They merge into a new algorithm,called POSLT（Partially Overlapped Sub-block Logarithmic Transformation）.POSLT exhibits good performance and controllability even if it has high computational complexity.Experimental results show its local adaptation for contrast enhancement and give firmly support to POSLT.

关 键 词：对比度增强 高动态范围图像 对数变换 对比度测度 子块 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]