基于DCP和色差校正的浑浊水体图像增强方法  

An Image Enhancement Method for Turbid Underwater Images Based on DCP and Color Restoration

作　　者：王浩[1,2] 唐晓凡 鲁亮 徐德南[1] 杨新 WANG Hao;TANG Xiao-fan;LU Liang;XU Den-an;YANG Xin(Huazhong University of Science and Technology,Wuhan Hubei 430070;Guangdong Intelligent Robotics Institute,Dongguan Guangdong 523808)

机构地区：[1]华中科技大学,湖北武汉430070 [2]广东省智能机器人研究院,广东东莞523808

出　　处：《数字技术与应用》2021年第2期115-117,共3页Digital Technology & Application

基　　金：东莞市引进创新创业领军人才计划资助;广东省项目资助(2019ZT08Z780)。

摘　　要：水下环境相对于空气环境,尤其是浑浊水体,成像条件更加恶劣,而对于很多水下工程,例如污水水下检测或者淡水湖泊鱼群定位,都需要良好的视觉条件才可以很好地完成任务。针对淡水湖泊、污染水域这类的浑浊水体,我们提出了一种基于DCP和色差校正的浑浊水体图像增强方法。该方法首先使用双边滤波器对图像进行去噪处理,然后根据暗通道先验(DCP)对图像进行去雾。最后在灰度世界理论的基础上结合对比度受限的自适应直方图均衡,对图像的对比度和色彩进行改善。为了验证算法的有效性,我们对浑浊水体的水下图像进行了处理,将处理结果与其他增强方法结果进行比较,同时使用水下评价标准对结果进行对比与分析。通过主观和客观两个方面的对比,验证该算法能够改善图像质量,使图像具有更好的色彩、对比度和更多的细节,这对水下作业具有重要意义。Underwater environment is harder than air,especially turbid underwater.For many underwater operations,such as the detection of polluted water or fish location of lake,good visual condition is essential.Focusing on improving image quality of these environment,we propose a novel image enhancement algorithm.Firstly,Bilateral Filter is used to reduce image noise,which is the basic for image dehazing based on Dark Channel Prior(DCP).Finally,we combine the Gray World theory with Contrast Limited Adaptive Histogram Equalization to improve the contrast and color of images.To verify the algorithm,the algorithm is used to process different turbid underwater images,which are also compared with results got by other enhancement methods.After getting result images,underwater image quality metrics are used to evaluate results,based on which we conclude that the proposed algorithm can improve the quality of images effectively.Result images are with better color,contrast and more details,which has a great significance to underwater operations.

关 键 词：图像增强 浑浊水体 暗通道先验去雾 颜色校正 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]