基于HSI的改进Retinex水下图像增强算法  

Improved Retinex Underwater Image Enhancement Algorithm Based on HSI Model

作　　者：贾芃 李博[1] 赵晓龙 JIA Peng;LI Bo;ZHAO Xiaolong(Key Laboratory of Instrumentation Science&Dynamic Measurement,Ministry of Education,North University of China,Taiyuan 030051,China)

机构地区：[1]中北大学仪器科学与动态测试教育部重点实验室,太原030051

出　　处：《实验室研究与探索》2020年第12期1-4,14,共5页Research and Exploration In Laboratory

基　　金：国家自然科学基金项目(61471325);山西省重点计划研发项目(201803D121061)。

摘　　要：针对水下采集的彩色图像会出现细节丢失、边缘模糊和颜色衰减等现象,将原始水下图像变换到HSI(Hue-Saturation-Intensity)色彩空间,在小波分解和多通道滤波基础上对传统Retinex算法进行了改进。利用小波变换将彩色图像分解为高频与低频部分,针对其不同频带上图像的特征,分别用多通道滤波和Retinex两种处理方法对图像进行处理。实验表明:该算法不仅能够有效消除光晕伪影现象,增强图像的亮度和对比度,同时也还原了图像原本的特征细节,并有效去除了图像噪声,图像的视觉效果与客观评价结果也都取得了较大提升。与传统的Retinex算法和直方图均衡法对比,该算法能够显著有效增强水下图像,主观和客观评价指标有大幅度的提升。In view of the loss of details,blurred edges and color attenuation of color images collected underwater,this paper transforms color images into HSI color space,improves Retinex algorithm by using wavelet decomposition and multichannel filtering,decomposes color images into high and low frequency parts by using wavelet transform,and processes the images with two algorithms,respectively.Experiments show that the algorithm not only effectively eliminates the halo artifacts,enhances the brightness and contrast of the image,but also restores the original feature details of the image,effectively removes image noise,and improves the visual effect and objective evaluation results of the image.Compared with the traditional Retinex algorithm and histogram equalization method,the proposed algorithm can significantly enhance underwater images,and the subjective and objective evaluation indexes can also be significantly improved.

关 键 词：水下图像 HSI模型 小波变换 多通道增强 

分 类 号：TN391.9[电子电信—物理电子学]