光照不均图像增强方法综述  

Survey on enhancement methods for non-uniform illumination image

作　　者：梁琳[1] 何卫平[1] 雷蕾[1] 张维[1] 王红霄[1] 

机构地区：[1]西北工业大学现代设计与集成制造重点实验室,西安710072

出　　处：《计算机应用研究》2010年第5期1625-1628,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(50505039);国防基础科研基金资助项目(B2720060292)

摘　　要：针对图像采集过程中因环境光照不佳或物体表面采光不均等原因造成的图像光照不均问题,分析了几类常用的增强方法,包括以直方图均衡化方法为代表的灰度变换法、基于照明—反射的同态滤波法、Retinex增强方法以及梯度域增强方法,对比了这几种方法的处理效果,指出了这些方法各自的适用范围。最后给出了光照不均问题解决方法的改进方向。In order to solve the non-uniform illumination problem of digital images, which was caused by poor light environment or lighting disproportionation on object’s surface, this paper provided some common enhancement methods.They included gray transformation represented by histogram equalization, homomorphic filtering based on illumination-reflectance model, Retinex enhancement theory and gradient field method.Compared these methods’ effect,analysed their application scopes and gave the solution direction of the non-uniform illumination problem.

关 键 词：光照不均 图像增强 直方图均衡 同态滤波 Retinex方法 梯度域 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]