Single underwater image enhancement by attenuation map guided color correction and detail preserved dehazing

被引:92
作者
Liang, Zheng [1 ]
Wang, Yafei [1 ]
Ding, Xueyan [1 ]
Mi, Zetian [1 ]
Fu, Xianping [1 ,2 ]
机构
[1] Dalian Maritime Univ, Informat Sci & Technol Sch, Dalian 116026, Peoples R China
[2] Pengcheng Lab, Shenzhen 518055, Guangdong, Peoples R China
基金
中国国家自然科学基金;
关键词
Underwater image enhancement; Color correction; Wavelength attenuation; Maximum intensity prior; Gradient domain;
D O I
10.1016/j.neucom.2020.03.091
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
The attenuation (sum of absorption and scattering), which caused by the dense and non-uniform medium, generally leads to problems of color degradation and detail loss in underwater imaging. To ad -dress these problems, we propose a systematic underwater image enhancement method, which includes an attenuation map guided underwater image color correction approach and a detail preserved dehazing approach. The color correction approach fully considers the main causes of color degradation in underwa-ter imaging, namely wavelength-dependent attenuation of different colors. According to the attenuation map of each color channel, a piece-wise linear transform is used to process the information of each color channel. Then, the detail preserved dehazing approach based on multi-scale decomposition is proposed to compensate for the lost details while eliminating the effects of haze. Especially, an adaptive Maximum Intensity Prior (MIP) measurement based on maximum attenuation identification is proposed to estimate transmission of the medium. Experiments on a variety types of degraded underwater images have proven that our proposed method can produce accurate results with vivid color and fine details, even better than other state-of-the-art underwater image dehazing methods. ? 2020 Published by Elsevier B.V.
引用
收藏
页码:160 / 172
页数:13
相关论文
共 53 条
[1]   A Revised Underwater Image Formation Model [J].
Akkaynak, Derya ;
Treibitz, Tali .
2018 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2018, :6723-6732
[2]   Color Balance and Fusion for Underwater Image Enhancement [J].
Ancuti, Codruta O. ;
Ancuti, Cosmin ;
De Vleeschouwer, Christophe ;
Bekaert, Philippe .
IEEE TRANSACTIONS ON IMAGE PROCESSING, 2018, 27 (01) :379-393
[3]   Locally Adaptive Color Correction for Underwater Image Dehazing and Matching [J].
Ancuti, Codruta O. ;
Ancuti, Cosmin ;
De Vleeschouwer, Christophe ;
Garcia, Rafael .
2017 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS (CVPRW), 2017, :997-1005
[4]  
Ancuti CO, 2011, IEEE IMAGE PROC, P1557, DOI 10.1109/ICIP.2011.6115744
[5]  
Ancuti C, 2016, INT C PATT RECOG, P4202, DOI 10.1109/ICPR.2016.7900293
[6]  
Ancuti C, 2012, PROC CVPR IEEE, P81, DOI 10.1109/CVPR.2012.6247661
[7]   Adaptive Stretching Method for Underwater Image Color Correction [J].
Ao, Jun ;
Ma, Chunbo .
INTERNATIONAL JOURNAL OF PATTERN RECOGNITION AND ARTIFICIAL INTELLIGENCE, 2018, 32 (02)
[8]  
Avidan S, 2017, P BRIT MACH VIS C BM, V1
[9]   A SPATIAL PROCESSOR MODEL FOR OBJECT COLOR-PERCEPTION [J].
BUCHSBAUM, G .
JOURNAL OF THE FRANKLIN INSTITUTE-ENGINEERING AND APPLIED MATHEMATICS, 1980, 310 (01) :1-26
[10]  
Carlevaris-Bianco N., 2010, OCEANS, P1, DOI DOI 10.1109/OCEANS.2010.5664428