2.765

2022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

基于并联卷积神经网络的图像去雾

陈清江 张雪

陈清江,  张雪.  基于并联卷积神经网络的图像去雾.  自动化学报,  2021,  47(7): 1739−1748 doi: 10.16383/j.aas.c190156
引用本文: 陈清江,  张雪.  基于并联卷积神经网络的图像去雾.  自动化学报,  2021,  47(7): 1739−1748 doi: 10.16383/j.aas.c190156
Chen Qing-Jiang,  Zhang Xue.  Single image dehazing based on multiple convolutional neural networks.  Acta Automatica Sinica,  2021,  47(7): 1739−1748 doi: 10.16383/j.aas.c190156
Citation: Chen Qing-Jiang,  Zhang Xue.  Single image dehazing based on multiple convolutional neural networks.  Acta Automatica Sinica,  2021,  47(7): 1739−1748 doi: 10.16383/j.aas.c190156

基于并联卷积神经网络的图像去雾

doi: 10.16383/j.aas.c190156
基金项目: 国家自然科学基金(61403298), 陕西省自然科学基金(2015JM1024), 陕西省教育厅专项科研计划(2013JK0586)资助
详细信息
    作者简介:

    陈清江:西安建筑科技大学理学院副教授. 2006年于西安交通大学计算数学专业获博士学位. 主要研究方向为小波分析与图像处理.E-mail: qjchen66xytu@126.com

    张雪:西安建筑科技大学理学院硕士研究生. 主要研究方向为小波分析与图像处理. 本文通信作者.E-mail: zhangxueyanice@163.com

Single Image Dehazing Based on Multiple Convolutional Neural Networks

Funds: Supported by National Natural Science Foundation of China (61403298), Natural Science Foundation of Shaanxi Province (2015JM1024), and Special Research Projects of Shaanxi Provincial Department of Education (2013JK0586)
More Information
    Author Bio:

    CHEN Qing-Jiang Associate professor at the College of Science, Xi' an University of Architecture and Technology. He received his Ph.D. degree in computational mathematics from Xi' an Jiaotong University in 2006. His research interest covers wavelet analysis and image processing

    ZHANG Xue Master student at the College of Science, Xi' an University of Architecture and Technology. Her research interest covers wavelet analysis and image processing. Corresponding author of this paper

  • 摘要:

    针对现有的单幅图像去雾问题, 提出了一种基于并联卷积神经网络的单幅图像去雾算法, 以端对端的方式实现图像去雾. 首先, 使用雾天RGB图像YUV变换的Y、U和V分量构建并联卷积神经网络, 自适应获得雾霾特征; 网络结构由两个子网络组成, 较深的网络预测清晰图像的亮度通道, 较浅的网络预测色度通道和饱和度通道. 最后, 采用递归双边滤波, 对去雾后的图像进行滤波, 可以得到更加清晰的无雾图像. 实验结果表明, 本文去雾算法无论是在合成雾天图像数据集还是自然雾天图像数据集上, 都具有良好的对比度与清晰度. 在主观评价和客观评价方面, 本文去雾算法都优于其他对比算法.

  • 图  1  本文网络结构

    Fig.  1  The network structure of this paper

    图  2  YUV颜色空间的各个分量对比图像

    Fig.  2  Each component of the contrast images of YUV color space

    图  3  本文算法流程图

    Fig.  3  Algorithm flow chart in this paper

    图  4  不同算法对雾天图像Dolls的去雾结果

    Fig.  4  Different algorithms to defog results of fog image Dolls

    图  6  不同算法对雾天图像Reindeer的去雾结果

    Fig.  6  Different algorithms to defog results of fog image Reindeer

    图  5  不同算法对雾天图像Trees的去雾结果

    Fig.  5  Different algorithms to defog results of fog image Trees

    图  7  不同算法的峰值信噪比和结构相似度的对比结果

    Fig.  7  Comparison results of PSNR and SSIM of different algorithms

    图  8  自然雾天图像House的去雾结果对比

    Fig.  8  Comparison of defogging results of natural foggy images House

    图  9  自然雾天图像Pumpkin的去雾结果对比

    Fig.  9  Comparison of defogging results of natural foggy images Pumpkin

    图  10  不同算法的对比度和平均梯度的对比结果

    Fig.  10  Comparison results of Contrast and average gradient of different algorithms

    表  1  多尺度卷积模型

    Table  1  Multi-scale convolution model

    卷积核尺寸补零步长
    1×1×1601
    3×3×16 1 1
    5×5×16 2 1
    7×7×1631
    下载: 导出CSV

    表  2  30幅合成雾天图像的RMSE, 色调还原度, 平均梯度, 信息熵, PSNR, SSIM的平均结果

    Table  2  The average results of RMSE, tone reduction, average gradient, information entropy, PSNR, SSIM for the 30 synthetic foggy images

    评价指标雾天图像 DCP CAP SRCNNDehazeNet MSCNN 本文算法
    RMSE $ \downarrow $0.15340.02890.03310.01790.01760.02410.0160
    色调还原度 $ \uparrow $0.53730.75230.65540.89000.89250.80640.9503
    平均梯度 $ \uparrow $5.19096.01406.01016.20066.03357.10387.2011
    信息熵 $ \uparrow $12.500414.585514.592016.600216.566016.485016.6522
    PSNR $ \uparrow $15.310219.864023.202125.230525.701725.689925.9057
    SSIM $ \uparrow $0.58830.85120.86660.93740.94610.86570.9603
    下载: 导出CSV

    表  3  图像House采用不同算法去雾后评价指标结果

    Table  3  Evaluation indicators results by different defogging algorithms for image House

    评价指标雾天图像DCPCAPSRCNNDehazeNetMSCNN本文算法
    均值 $ \uparrow $85.263085.562086.150890.400789.470983.443793.6325
    标准差 $ \uparrow $32.594323.022128.568133.926137.951256.165259.3654
    信息熵 $ \uparrow $15.222016.423815.929416.810415.873814.162416.8542
    下载: 导出CSV

    表  4  图像Pumpkin采用不同算法去雾后评价指标结果

    Table  4  Evaluation indicators results by different defogging algorithms for image Pumpkin

    评价指标雾天图像DCPCAPSRCNNDehazeNetMSCNN本文算法
    均值 $ \uparrow $79.032179.233588.826277.843697.192494.246897.9533
    标准差 $ \uparrow $45.996947.447944.814544.802050.428243.165859.6355
    信息熵 $ \uparrow $14.895414.325515.477815.999215.976115.978316.7523
    下载: 导出CSV

    表  5  不同算法的运行时间对比结果 (s)

    Table  5  The run time comparison results of different algorithms (s)

    图像像素尺寸DCPCAPSRCNNDehazeNetMSCNN本文算法
    House345×4501.0269851.0830501.4200000.7127002.3000000.386680
    Pumpkin600×4501.0408804.0978662.4135270.6433542.6328090.401389
    下载: 导出CSV
  • [1] McCann J J. Retinex at 40. Journal Electronic Imaging, 2004, 13(1): 6−7 doi: 10.1117/1.1645250
    [2] Cooper T J, Baqai F A. Analysis and extensions of the Frankle-McCann retinex algorithm. Journal of Electronic Imaging, 2004, 13(1): 85−92 doi: 10.1117/1.1636182
    [3] Land E H, McCann J J. Lightness and retinex theory. Journal of the Optical Society of America, 1971, 61(1): 1−11 doi: 10.1364/JOSA.61.000001
    [4] Land E H. An alternative technique for the computation of the designator in the retinex theory of color vision. Proceedings of the National Academy of Sciences of the United States of America, 1986, 83(10): 3078−3080 doi: 10.1073/pnas.83.10.3078
    [5] Frankle J A, McCann J J. Method and Apparatus for Lightness Imaging: USA. Patent 4384336, May 1983.
    [6] 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
    [7] Zhu Q S, Mai J M, Shao L. A fast single image haze removal algorithm using color attenuation prior. IEEE Transactions on Image Processing, 2015, 24(11): 3522−3533 doi: 10.1109/TIP.2015.2446191
    [8] 周全, 王磊, 周亮, 郑宝玉. 多尺度上下文的图像标注算法. 自动化学报, 2014, 40(12): 2944−2949

    Zhou Quan, Wang Lei, Zhou Liang, Zheng Bao-Yu. Multi-scale contextual image labeling. Acta Automatica Sinica, 2014, 40(12): 2944−2949
    [9] 李新德, 杨伟东, Dezert Jean. 一种飞机图像目标多特征信息融合识别方法. 自动化学报, 2012, 38(8): 1298−1307

    Li Xin-De, Yang Wei-Dong, Dezert Jean. An airplane image target′s multi-feature fusion recognition method. Acta Automatica Sinica, 2012, 38(8): 1298−1307
    [10] 胡长胜, 詹曙, 吴从中. 基于深度特征学习的图像超分辨率重建. 自动化学报, 2017, 43(5): 814−821

    Hu Chang-Sheng, Zhan Shu, Wu Cong-Zhong. Image super-resolution based on deep learning features. Acta Automatica Sinica, 2017, 43(5): 814−821
    [11] Cai B L, Xu X M, Jia K, Qing C M, Tao D C. DehazeNet: An end-to-end system for single image haze removal. IEEE Transactions on Image Processing, 2016, 25(15): 5187−5198
    [12] Ren W Q, Liu S, Zhang H, Pan J S, Cao X C, Yang M H. Single image dehazing via multi-scale convolutional neural networks. In: Proceedings of the 2016 Conference on Computer Vision–ECCV2016, 2016. 154−169
    [13] Li C Y, Guo J C, Porikli F, Fu H Z, Pan Y W. A cascaded convolutional neural network for single image dehazing. IEEE Access, 2018, 6: 24877−24887
    [14] Narasimhan S G, Nayar S K. Contrast restoration of weather degraded images. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2003, 25(6): 713−724
    [15] 徐岩, 孙美双. 基于多特征融合的卷积神经网络图像去雾算法. 激光与光电子学进展, 2018, 55(3): 260−269

    Xu Yan, Sun Mei-Shuang. Convolution neural network image defogging based on multi-feature fusion. Laser and Optoelectronics Progress, 2018, 55(3): 260−269
    [16] Dong C, Chen C L, He K M, Tang X O. Image super-resolution using deep convolutional networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2016, 38(2): 295−307 doi: 10.1109/TPAMI.2015.2439281
    [17] 李大鹏, 禹晶, 肖创柏. 图像去雾的无参考客观质量评测方法. 中国图象图形学报, 2011, 16(9): 1753−1757 doi: 10.11834/jig.110024

    Li Da-Peng, Yu Jing, Xiao Chuang-Bai. No-reference quality assessment method for defogged images. Journal of Image and Graphics, 2011, 16(9): 1753−1757 doi: 10.11834/jig.110024
  • 加载中
图(10) / 表(5)
计量
  • 文章访问数:  1371
  • HTML全文浏览量:  559
  • PDF下载量:  318
  • 被引次数: 0
出版历程
  • 收稿日期:  2019-03-13
  • 录用日期:  2019-10-11
  • 刊出日期:  2021-07-27

目录

    /

    返回文章
    返回