基于幂次变换和MSR的光照不均图像增强  

Non-uniform Illumination Image Enhancement Based on the Power Low Transformation and MSR

作　　者：张亚飞[1] ZHANG Ya-fei （Department of Communication Engineering, Nan jing Institute of Technology, Nanjing 211167, China ）

机构地区：[1]南京工程学院通信工程学院,江苏南京211167

出　　处：《电脑知识与技术》2012年第8期5456-5458,共3页Computer Knowledge and Technology

基　　金：南京工程学院科研基金项目（YKJ200907）

摘　　要：针对光照不均图像暗区细节丢失的问题，提出了基于幂次变换和MSR（多尺度Retinex）的光照不均图像增强算法。首先对光照不均图像进行幂次变换，扩展图像暗区灰度值的动态范围，增强暗区的细节，然后再利用MSR算法对幂次变换的结果进行进一步处理，提高图像的对比度实验表明，该算法在消除不均匀光照、保持图像较高的对比度的同时，有效地增强了原始图像中暗区的细节。Concerning the loss of details in the dark areas of non-uniform illumination images, the algorithm of non-uniform illumination image enhancement based on the power transform and MSR（Muhi-Scale Retinex） is proposed. Firstly, the power low transformation on non-uniform illumination image is used to expand the dynamic range of the gray value in the dark areas and enhance the image details in the dark areas. Then, the result of the power low transformation is processed further based on MSP,- to improve the image contrast. The ex- periments show that the algorithm can eliminate non-uniform illumination and maintain the high contrast of the image, at the same time, enhance the details of dark areas in the original image.

关 键 词：光照不均图像 图像增强 幂次变换 多尺度RETINEX 

分 类 号：TP391[自动化与计算机技术—计算机应用技术;自动化与计算机技术—计算机科学与技术]