Optics and Precision Engineering, Volume. 31, Issue 12, 1827(2023)

Image dehazing based on polarization optimization and atmosphere light correction

Jing WU1...2, Wenjie SONG1,2, Cuixia GUO1,2, Xiaojing YE1,2, and Feng HUANG12,* |Show fewer author(s)
Author Affiliations
  • 1College of Mechanical Engineering and Automation, Fuzhou University, Fuzhou3506, China
  • 2Institute of Advanced Technology Innovation, Fuzhou University, Fuzhou350116, China
  • show less
    Figures & Tables(18)
    Framework of proposed algorithm
    Optimized degree of polarization for fog images
    Comparison of degrees of polarization deviation of light and dark sides of fog-free and foggy scenes
    Comparison of dehazed images using different degree of polarization
    Comparison of dehazed images via enlarged or reduced degree of polarization
    Optimized degree of polarization and atmosphere light image obtained by different degree of polarization
    Foggy images captured in different density of fog
    Schematic diagram of LUCID camera acquisition of polarized state images
    Range of mean intensity in atmosphere light of foggy images captured in different density of fog
    Dehazed influence caused by different intensity of AO
    Results with or without atmospheric light correction
    Comparison of the results of algorithm in this paper
    Comparison of results in hazy scenes with sky from different algorithms
    Comparison of results in hazy scenes without sky region from different algorithms
    Comparison of results in hazy scene in distant hazy scenes from different algorithms
    Dehazing results of different algorithms in fog scenarios with different concentrations. The top, middle, and bottom are mist, medium fog, and dense fog scenes
    • Table 1. Average objective evaluation metric on the polarization dataset without reference

      View table
      View in Article

      Table 1. Average objective evaluation metric on the polarization dataset without reference

      MethodContrastEntropySTD
      基准(S024.296 26.677 240.672 9
      GDCP32.755 76.345 933.610 4
      AodNet57.956 06.214 841.320 0
      MSCNN39.492 96.658 341.148 0
      MBEM156.569 16.469 541.941 1
      CAD70.777 26.412 337.206 5
      4K36.498 76.886 340.493 9
      PSD150.318 97.109 859.518 9
      D449.505 46.413 848.174 6
      SFDF190.405 96.058 451.907 0
      GPLF90.391 07.091 247.584 3
      Proposed98.894 67.292 565.832 3
    • Table 2. Average objective evaluation metric on the polarization dataset with reference

      View table
      View in Article

      Table 2. Average objective evaluation metric on the polarization dataset with reference

      IndexGDCPAodNetMSCNNMBEMCADMOF4KPSDD4SFDFGPLFProposed
      MistSSIM0.737 60.828 10.853 60.827 20.726 80.698 20.861 30.800 90.601 10.812 70.846 30.868 8
      PSNR11.638 715.474 217.973 718.472 412.075 616.524 418.203 818.162 110.179 216.575 617.825 818.118 7
      CIEDE22.947 714.498 511.803 611.822 120.707 818.354 610.320 911.558 825.128 711.925 811.239 811.714 2

      Mediu-m

      fog

      SSIM0.704 60.813 70.828 00.797 90.733 20.576 50.812 80.780 60.703 30.771 80.809 50.832 7
      PSNR11.284 216.053 717.991 118.428 612.860 215.161 416.135 116.935 312.898 014.381 917.190 418.207 2
      CIEDE23.044 314.082 911.675 512.253 619.320 420.436 713.860 313.343 319.884 715.231 512.249 912.011 4

      Dense

      fog

      SSIM0.739 00.781 00.761 30.718 80.697 70.375 40.741 90.715 20.757 40.715 80.727 90.764 6
      PSNR13.341 316.023 515.419 116.090 815.137 212.614 713.719 413.279 814.937 211.598 215.432 215.905 0
      CIEDE18.393 614.322 514.277 914.570 216.410 625.370 218.303 518.170 315.004 021.376 414.886 713.549 2
      AverageSSIM0.727 10.807 60.814 30.781 30.719 20.550 00.805 30.765 60.687 30.766 70.794 60.822 0
      PSNR12.087 515.850 517.128 017.663 913.357 714.766 816.019 416.125 712.671 514.185 216.816 117.410 3
      CIEDE21.461 914.301 312.585 712.882 018.812 921.387 214.161 614.357 520.005 816.177 912.792 112.424 9
    Tools

    Get Citation

    Copy Citation Text

    Jing WU, Wenjie SONG, Cuixia GUO, Xiaojing YE, Feng HUANG. Image dehazing based on polarization optimization and atmosphere light correction[J]. Optics and Precision Engineering, 2023, 31(12): 1827

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category: Information Sciences

    Received: Nov. 5, 2022

    Accepted: --

    Published Online: Jul. 25, 2023

    The Author Email: HUANG Feng (huangf@fzu.edu.en)

    DOI:10.37188/OPE.20233112.1827

    Topics