基于非锐化掩模引导滤波的水下图像细节增强算法研究  

Research on underwater image detail enhancement based on unsharp mask guided filtering

作　　者：付青青[1] 景春雷 裴彦良 阚光明 张正炳[1] 吴爱平[1] Fu Qingqing;Jing Chunlei;Pei Yanliang;Kan Guangming;Zhang Zhengbing;Wu Aiping(Electronics&Information School,Yangtze University,Jingzhou 434023,China;National Deep Sea Center,Qingdao 266237,China;First Institute of Oceanography,Ministry of Natural Resources,Qingdao 266061,China;Key Laboratory of Marine Sedimentology and Environmental Geology,Ministry of Natural Resources,Qingdao 266061,China;Laboratory for Marine Geology and Environment,Pilot National Laboratory for Marine Science and Technology(Qingdao),Qingdao 266237,China)

机构地区：[1]长江大学电子信息学院,湖北荆州434023 [2]国家深海基地管理中心,山东青岛266237 [3]自然资源部第一海洋研究所,山东青岛266061 [4]海洋沉积与环境地质自然资源部重点实验室,山东青岛266061 [5]青岛海洋科学与技术试点国家实验室海洋地质过程与环境功能实验室,山东青岛266237

出　　处：《海洋学报》2020年第7期130-138,共9页

基　　金：国家自然科学基金(51604038);青岛海洋科学与技术试点国家实验室海洋地质过程与环境功能实验室开放基金(MGQNLMKF201705);长江大学优秀博士、硕士学位论文培育计划资助项目。

摘　　要：针对水下图像对比度偏低,细节模糊的问题,本文提出基于非锐化掩模引导滤波的细节增强方法。首先由原始图像做引导图进行滤波得到细节层图像,并对细节层使用噪声检测的中值滤波去除斑点噪声;然后对原始图像进行基于均值滤波的非锐化掩模,得到锐化图像,并将锐化图像作为引导图对原始图像进行引导滤波,获取基础层图像;最后将滤波后的细节层进行增益后与引导滤波获取的基础层进行叠加,达到增强水下图像细节的目的。并通过信息熵、局部对比度和平均梯度3种客观评价指标对图像处理结果进行了对比分析,主观和客观测试结果表明,本文采用的算法能够有效提高图像对比度以及增强细节信息,有利于提高水下图像资料解释的准确性。To overcome the problem of the low contrast and blurred details of underwater images,we propose a method to enhance the details of underwater images based on unsharp mask guided filtering in this paper.Firstly,the original image is guided filtering to obtain a detail layer image,and the speckle noise in detail layer image is removed by the median filtering based on noise detection.Then the unsharp mask based on mean filtering is operated on the original image to obtain a sharpened image,which is used as guidance image to perform guided filtering on the original image,and the base layer image is obtained.It is magnified to boost the details so that the details of the output image are obtained.Finally,the filtered detail layer is magnified to boost the details,and the enhanced underwater image is the combination of the boosted detail layer and the base layer.The image processing results are compared and analyzed through three objective evaluation indexes:information entropy,local contrast and average gradient.Subjective and objective test results show that the proposed algorithm can effectively improve image contrast and detail information,which is beneficial to improve the accuracy of underwater image data interpretation.

关 键 词：水下图像 引导滤波 细节增强 非锐化掩模 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]