水下彩色图像的亮度通道多尺度Retinex增强算法  

Multi-Scale Retinex Enhancement Algorithm on Luminance Channel of Color Underwater Image

作　　者：张凯[1] 金伟其 裘溯[1] 王霞[1] 

机构地区：[1]北京理工大学光电学院"光电成像技术与系统"教育部重点实验室,北京100081

出　　处：《红外技术》2011年第11期630-634,共5页Infrared Technology

基　　金：国家863计划资助项目(2006AA09Z207)

摘　　要：水体散射使得水下光电成像所获得的图像存在较为均匀的散射光分布,降低了水下图像的清晰度和可观察景深,减小水体散射光的影响是提高水下光电成像质量的重要途径。论文基于Retinex理论,将水体散射效应等效为环境光照的变化,通过水下彩色图像亮度通道下的多尺度Retinex(MSR)算法处理,有效地减小水体散射效应,提高所获图像中水下目标的真实性。实验表明,MSR算法能对含有不同程度水体散射效应的水下彩色图像进行全局视觉效果提升,更清晰地突出物体的细节并获得较鲜艳的色彩表现。Water scattering effect make images taken by underwater optoelectronic system casting with a uniform scattering distribution, which decrease the clarity and visible distance of underwater images. Reducing this effect is an important way to improve the quality of underwater imaging. In this article, we treat water scattering effect as changes of environmental luminance. By MSR calculations for luminance channel of color underwater image, we can effectively reduce scattering of water and increase true characteristic of underwater objects. Experimental results show that, our MSR algorithm on luminance channel of color underwater images with different scattering degrees is able to rise up their global visual effects, make details more visible and achieve better color reoresentation.

关 键 词：水下彩色图像 散射效应 RETINEX MSR 图像增强 

分 类 号：TN911.73[电子电信—通信与信息系统]