一种自适应SSR的雾天低照度图像增强方法  

AN ADAPTIVE SSR METHOD FOR FOGGY LOW ILLUMINATION IMAGE ENHANCEMENT

作　　者：李灿林 朱金娟 刘金华 毕丽华 Li Canlin;Zhu Jinjuan;Liu Jinhua;Bi Lihua(College of Computer and Communication Engineering,Zhengzhou University of Light Industry,Zhengzhou 450000,Henan,China;Office of Discipline Construction,Zhengzhou University of Light Industry,Zhengzhou 450000,Henan,China)

机构地区：[1]郑州轻工业大学计算机与通信工程学院,河南郑州450000 [2]郑州轻工业大学学科建设办公室,河南郑州450000

出　　处：《计算机应用与软件》2022年第9期233-239,268,共8页Computer Applications and Software

基　　金：国家自然科学基金项目(U1404623);河南省科技攻关项目(152102210350)。

摘　　要：针对雾天低照度情况下图像内容模糊、对比度低等的图像质量退化现象,提出一种自适应单尺度Retinex(SSR)的雾天低照度图像增强方法。该方法将图像转换到HSI空间,对其亮度分量进行伽马自适应校正;结合图像信息熵和灰度标准差设计评价函数,采用粒子群优化算法(Particle Swarm Optimization,PSO)自适应地选取SSR算法中中心环绕函数的参数值,对图像进行处理。通过实验,与几种算法进行比较,表明该方法能很好地改善传统SSR算法的缺陷,且在提升图像清晰度和对比度上的效果明显。In view of the image quality degradation,for instance,blurred content and low contrast of the image in low ilumination in foggy days,this paper proposes an adaptive single-scale Retinex(SSR)method for foggy low illumination image enhancement.It converted the image to HSI space and perform Gamma adaptive correction on its luminance component.The image information entropy and standard deviation was combined to design an evaluation function,and the parameter values of the center surround function was selected in the SSR algorithm adaptively by particle swarm optimization(PSO)to process the image.Compared with several algorithms by experiments,it shows that this method can improve the defects of traditional SSR algorithm,and its effect on improving image sharpness and contrast is obvious.

关 键 词：雾天低照度 单尺度Retinex算法 HSI空间 伽马校正 粒子群优化算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]