基于自动色阶和多尺度Retinex彩色图像增强算法  

Color Image Enhancement Algorithm Based on Multi-scale Retinex and Automatic Color Method

作　　者：冯红波 李萍 李波 FENG Hongbo;LI Ping;LI Bo(School of physics and electronic and electrical engineering,Ningxia University,Yinchuan 750021,China)

机构地区：[1]宁夏大学物理与电子电气工程学院

出　　处：《无线电工程》2019年第10期910-914,共5页Radio Engineering

基　　金：2018年宁夏高等学校科学研究项目(NGY2018012)

摘　　要：为了解决多尺度Retinex算法(MSR)在增强光照不均的单彩色图像时出现的细节模糊、颜色失真的问题,提出一种基于自动色阶和多尺度Retinex算法(Auto-colorMSR)。对图像进行双边滤波去噪,利用添加光照参数调节的MSR算法对去噪图像进行增强,并通过自动色阶进行色彩恢复得到增强图像。实验结果表明,提出的算法优于传统算法,图像增强后具有更好颜色保真和边缘细节保留特性。In order to solve the problems such as detail blur and color distortion caused by multi-scale retinex algorithm(MSR) in enhancing single-color image with uneven illumination,an automatic color scale and multi-scale retinex algorithm(auto-colorMSR) is proposed.Firstly,the image is double-filtered and denoised,and the denoised image is enhanced by the MSR algorithm with the added illumination parameter adjustment.Finally,the color image is restored by the automatic color gradation to obtain the enhanced image.The experimental results show that the proposed algorithm outperforms the traditional algorithm and has better color fidelity and edge detail retention after image enhancement.

关 键 词：图像增强 RETINEX算法 彩恢复 自动色阶 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]