2.793

2018影响因子

(CJCR)

  • 中文核心
  • EI
  • 中国科技核心
  • Scopus
  • CSCD
  • 英国科学文摘

留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

基于改进CycleGAN的水下图像颜色校正与增强

李庆忠 白文秀 牛炯

李庆忠, 白文秀, 牛炯. 基于改进CycleGAN的水下图像颜色校正与增强. 自动化学报, 2020, 46(x): 1−11 doi: 10.16383/j.aas.c200510
引用本文: 李庆忠, 白文秀, 牛炯. 基于改进CycleGAN的水下图像颜色校正与增强. 自动化学报, 2020, 46(x): 1−11 doi: 10.16383/j.aas.c200510
Li Qing-Zhong, Bai Wen-Xiu, Niu Jiong. Underwater image color correction and enhancement based on improved cycle-consistent generative adversarial networks. Acta Automatica Sinica, 2020, 46(x): 1−11 doi: 10.16383/j.aas.c200510
Citation: Li Qing-Zhong, Bai Wen-Xiu, Niu Jiong. Underwater image color correction and enhancement based on improved cycle-consistent generative adversarial networks. Acta Automatica Sinica, 2020, 46(x): 1−11 doi: 10.16383/j.aas.c200510

基于改进CycleGAN的水下图像颜色校正与增强

doi: 10.16383/j.aas.c200510
基金项目: 国家重点研发计划(2017YFC1405202), 海洋公益性行业科研专项(201605002)资助
详细信息
    作者简介:

    李庆忠:于2000年于中国农业大学获得博士学位, 现为中国海洋大学工程学院教授, 主要从事信号处理, 图像处理与模式识别等方面的研究. E-mail: liqingzhong@ouc.edu.cn

    白文秀:中国海洋大学硕士研究生, 主要研究方向为图像处理与模式识别. 本文通信作者. E-mail: baiwenxiu@ouc.edu.cn

    牛炯:博士研究生, 工程师, 主要研究方向为高频雷达信号处理、高频雷达海洋环境监测技术. E-mail: niujiong@ouc.edu.cn

Underwater Image Color Correction and Enhancement Based on iMproved Cycle-Consistent Generative Adversarial Networks

Funds: Supported by National Key Research and Development Program of China (2017YFC1405202), National Marine Technology Program for Public Welfare of China (No.201605002)
  • 摘要: 针对水下观测图像的颜色失真和散射模糊问题, 提出一种基于改进循环一致性生成对抗网络(Cycle-consistent generative adversarial networks, CycleGAN)的水下图像颜色校正与增强算法. 为了利用CycleGAN学习水下降质图像到空气中图像的映射关系, 对传统CycleGAN的损失函数进行了改进, 提出了基于图像强边缘结构相似度(Strong edge and structure similarity, SESS)损失函数的SESS-CycleGAN, SESS-CycleGAN可以在保留原水下图像的边缘结构信息的前提下实现水下降质图像的颜色校正和对比度增强. 为了确保增强后图像和真实脱水图像颜色的一致性, 建立了SESS-CycleGAN和正向生成网络G相结合的网络结构; 并提出了两阶段学习策略, 即先利用非成对训练集以弱监督方式进行SESS-CycleGAN学习, 然后再利用少量成对训练集以强监督方式进行正向生成网络G的监督式学习. 实验结果表明: 本文算法在校正水下图像颜色失真的同时还增强了图像对比度, 且较好地实现了增强后图像和真实脱水图像视觉颜色的一致性.
  • 图  1  双级网络结构

    Fig.  1  Dual-level network structure

    图  2  生成器结构图

    Fig.  2  Structure of generation network

    图  3  判别器网络结构

    Fig.  3  Structure of discriminator

    图  4  生成网络特征流程图

    Fig.  4  Feature flow chart of generation network

    图  5  强边缘图像对比

    Fig.  5  Comparison of strong edge images

    图  6  SESS-CycleGAN损失函数曲线

    Fig.  6  Loss function curve of SESS-CycleGAN

    图  7  强监督网络损失函数曲线

    Fig.  7  Loss function curve of SESS-CycleGAN

    图  8  颜色失真图像增强结果对比

    Fig.  8  Comparison of enhanced results for color distortion images

    图  9  颜色失真图像增强结果对比

    Fig.  9  Comparison of enhanced results for color distortion images

    图  10  增强后图像与真实脱水图像对比

    Fig.  10  Comparison of enhanced images with real air images

    表  1  增强图像的指标对比

    Table  1  Comparison of enhanced images

    Method Entropy UCIQE PSNR
    Original 6.546 0.437
    DCP 6.940 0.540 19.537
    GW 6.621 0.493 16.491
    CycleGAN 7.592 0.613 19.690
    SSIM-CycleGAN 7.598 0.541 20.670
    Ours 7.714 0.631 24.594
    下载: 导出CSV
  • [1] Pinto A M, Matos, Anibal C. M. A hybrid imaging system for underwater robot ic applications. Information Fusion, 2020, 55: 16−29 doi: 10.1016/j.inffus.2019.07.014
    [2] Chadebecq F, Vasconcelos F, Lacher R. Refractive Two-View Recon- struction for Underwater 3D Vision. International Journal of Computer Vision, 2020, 128(5): 1101−1117 doi: 10.1007/s11263-019-01218-9
    [3] Gallimore E, Terrill E, Pietruszka A, Gee J, Nager A, Hess R. Magnetic survey and autonomous target reacquisition with a scalar magnetometer on a small AUV. Journal of field robotics, 2020, 37(7): 1246−1266
    [4] Ahn J, Yasukawa S, Sonoda T. An Optical Image Transmission System for Deep Sea Creature Sampling Missions Using Autonomous Underwater Vehicle. IEEE Journal of oceanic Eengineering, 2020, 45(2): 350−361 doi: 10.1109/JOE.2018.2872500
    [5] He K M, Sun J, Tang X O. Single image haze removal using dark channel prior. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(12): 2341−2353 doi: 10.1109/TPAMI.2010.168
    [6] Wen H C, Tian Y H, Huang T J, Gao W. Single underwater image enhancement with a new optical model. In: IEEE International Symposium on Circuits and Systems(ISC 2013), Beijing. New York: IEEE, 2013: 753−756.
    [7] 杨爱萍, 曲畅, 王建, 张莉云. 基于水下成像模型的图像清晰化算法. 电子与信息学报, 2018, 40(2): 298−305

    Yang A P, Qu C, Wang J, Zhang L Y. Underwater image visibility restoration based on underwater imaging model. Journal of Electronics & Information Technology, 2018, 40(2): 298−305
    [8] Li C Y, Guo J C. Underwater image enhancement by dehazing and color correction. Journal of Electronic Imaging, 2015, 24(3): #033023
    [9] Emberton S, Chittka L, Cavallaro A. Underwater image and video dehazing with pure haze region segmentation. Computer Vision and Image Understanding, 2018, 168: 145−156 doi: 10.1016/j.cviu.2017.08.003
    [10] Liu Y C, Chan W H, Chen Y Q. Automatic white balance for digital still camera. IEEE Transactions on Consumer Electronics, 1995, 41(3): 460−466 doi: 10.1109/30.468045
    [11] Ebner M. Color Constancy. New York: John Wiley & Sons, 2007.
    [12] Iqbal K, Odetayo M O, James A E, Salam R A, Talib A Z H. Enhancing the low quality images using Unsupervised Colour Correction Method. In: Proceedings of the IEEE International Conference on Systems, Man and Cybernetics, Istanbul, Turkey, IEEE, 2010. 1703−1709.
    [13] Hitam M S, Awalludin E A, Yussof W N J H W, Bachok Z. Mixture contrast limited adaptive histogram equalization for underwater image enhancement. In: International Conference on Computer Applications Technology (ICCAT). Sousse, Tunisia: IEEE, 2013. 1−5.
    [14] Goodfellow I J, Pouget-Abadie J, Mirza M, Xu B, Bengio Y. Generative adversarial nets. In: Conference on Neural Information Processing Systems (NIPS), Montreal, CANADA. 2014. 2672−2680.
    [15] Li, J, Skinner K A, Eustice R M, Johnson-Roberson M. WaterGAN: Unsupervised generative network to enable realtime color correction of monocular underwater images. IEEE Robotics and Automation Letters, 2017, 3(1): 387−394
    [16] Li C Y, Guo J C, Guo C L. Emerging from Water: Underwater image color correction based on weakly supervised color transfer. IEEE Signal Processing Letters, 2018, 25(3): 323−327 doi: 10.1109/LSP.2018.2792050
    [17] Y Zhu, T Park, P Isola, A Efros. Unpaired image-to-image translation using cycle-consistent adversarial networks. In: IEEE International. Conference on Computer Vision (ICCV), Venice, ITALY. 2017. 2242−2251.
    [18] Chongyi L, Chunle G, Wenqi R. An Underwater Image Enhancement Benchmark Dataset and Beyond. IEEE Transactions on Image Processing, 2019. 29: 4376−4389.
    [19] Ahmad S, Abdul G, Nor A M I. Underwater image quality enhancement through integrated color model with Rayleigh distribution. Applied Soft Computing, 2015, 27: 219−230 doi: 10.1016/j.asoc.2014.11.020
    [20] Yang M, Sowmya A. An underwater color image quality evaluation metric. IEEE Transactions on Image Processing, 2015, 24(12): 6062−6071 doi: 10.1109/TIP.2015.2491020
    [21] Chen B, Zhang X, Wang R T, Li Z, Deng W. Detect concrete cracks based on OTSU algorithm with differential image. The Journal of Engineering, 2019, 23: 9088−9091
  • 加载中
计量
  • 文章访问数:  380
  • HTML全文浏览量:  146
  • 被引次数: 0
出版历程
  • 收稿日期:  2020-07-09
  • 修回日期:  2020-11-04
  • 网络出版日期:  2020-12-08

目录

    /

    返回文章
    返回