Journals Articles Conferences News About CLP
Submit a paper Email Alert
Search
Search
Articles
Journals
News
Advanced Search
Top Searches
2D MaterialsTransformation opticsQuantum PhotonicsSmall lasersSemiconductor UV PhotonicsSupersymmetric microring laser arrays

Acta Optica Sinica, Volume. 42, Issue 4, 0410002(2022)

Rapid Deep-Sea Image Restoration Algorithm Applied to Unmanned Underwater Vehicles

Wei Guo1,3, Youbo Zhang1,2、*, Yue Zhou2, Gaofei Xu1, and Guangwei Li1
Author Affiliations
  • 1Institute of Deep-Sea Science and Engineering, Chinese Academy of Sciences, Sanya, Hainan 572000, China
  • 2College of Engineering Science and Technology, Shanghai Ocean University, Shanghai 201306, China
  • 3University of Chinese Academy of Sciences, Beijing 100049, China
    • show less
    AbstractGet PDF(in Chinese)
    Figures&Tables (19)
    Equations (0)
    References (22)
    Cited By (2)
    Tools
    Paper Information
    Topics
    Figures & Tables(19)
    Examples of deep-sea images

    Fig. 1. Examples of deep-sea images

    Download full size
    Schematic of deep-sea imaging

    Fig. 2. Schematic of deep-sea imaging

    Download full size
    Examples of deep-sea image restoration under different algorithms. (a) Original image; (b) DCP algorithm; (c) UDCP algorithm; (d) MIP algorithm

    Fig. 3. Examples of deep-sea image restoration under different algorithms. (a) Original image; (b) DCP algorithm; (c) UDCP algorithm; (d) MIP algorithm

    Download full size
    Correlation study of depth of field in underwater scene. (a) Depth of field division of original deep-sea image; (b) statistical results of near, middle, and far scene related indicators of deep-sea images

    Fig. 4. Correlation study of depth of field in underwater scene. (a) Depth of field division of original deep-sea image; (b) statistical results of near, middle, and far scene related indicators of deep-sea images

    Download full size
    Depth of field correlation index statistics of deep-sea image scenes. (a) Saturation; (b) brightness; (c) difference between red and maximum blue-green channels

    Fig. 5. Depth of field correlation index statistics of deep-sea image scenes. (a) Saturation; (b) brightness; (c) difference between red and maximum blue-green channels

    Download full size
    Process of generating training sample lables for deep-sea images. (a) Original image; (b) preliminary depth map; (c) depth map after restoration

    Fig. 6. Process of generating training sample lables for deep-sea images. (a) Original image; (b) preliminary depth map; (c) depth map after restoration

    Download full size
    Background light estimation based on depth of field. (a) Background light estimated by maximum value; (b) background light estimated by the first 0.1% maximum value

    Fig. 7. Background light estimation based on depth of field. (a) Background light estimated by maximum value; (b) background light estimated by the first 0.1% maximum value

    Download full size
    Process of deep-sea image enhancement algorithm

    Fig. 8. Process of deep-sea image enhancement algorithm

    Download full size
    Enhancement results of different algorithms. (a) Original images; (b) UDCP algorithm; (c) MIP algorithm; (d) Ref. [12]; (e) Ref. [13]; (f) proposed algorithm

    Fig. 9. Enhancement results of different algorithms. (a) Original images; (b) UDCP algorithm; (c) MIP algorithm; (d) Ref. [12]; (e) Ref. [13]; (f) proposed algorithm

    Download full size
    Img 1 image magnification comparison. (a) Original images; (b) UDCP algorithm; (c) MIP algorithm; (d) Ref. [12]; (e) Ref. [13]; (f) proposed algorithm

    Fig. 10. Img 1 image magnification comparison. (a) Original images; (b) UDCP algorithm; (c) MIP algorithm; (d) Ref. [12]; (e) Ref. [13]; (f) proposed algorithm

    Download full size
    Img 4 image magnification comparison. (a) Original images; (b) UDCP algorithm; (c) MIP algorithm; (d) Ref. [12]; (e) Ref. [13]; (f) proposed algorithm

    Fig. 11. Img 4 image magnification comparison. (a) Original images; (b) UDCP algorithm; (c) MIP algorithm; (d) Ref. [12]; (e) Ref. [13]; (f) proposed algorithm

    Download full size
    Matching results of feature points before and after enhancement

    Fig. 12. Matching results of feature points before and after enhancement

    Download full size
    Feature point matching results enhanced by different algorithms. (a) DCP algorithm; (b) MIP algorithm

    Fig. 13. Feature point matching results enhanced by different algorithms. (a) DCP algorithm; (b) MIP algorithm

    Download full size
    Feature point matching results enhanced by different algorithms. (a) Ref. [12]; (b) Ref. [13]

    Fig. 14. Feature point matching results enhanced by different algorithms. (a) Ref. [12]; (b) Ref. [13]

    Download full size

    • Table 1. Coefficient of depth of field model

      View table

      Table 1. Coefficient of depth of field model

      Coefficientu0u1u2u3
      Value2.8621301250.401662315-3.565827900-1.535251780

    • Table 2. Comparison of number of feature points matching after image enhancement

      View table

      Table 2. Comparison of number of feature points matching after image enhancement

      ImgOriginal imageEnhanced by proposed algorithm
      1517
      215
      3413
      4916

    • Table 3. Number of feature points matching after image enhancement by contrast algorithm

      View table

      Table 3. Number of feature points matching after image enhancement by contrast algorithm

      ImgDCPalgorithmMIPalgorithmRef. [12]Ref. [13]
      110354
      21341
      37422
      410255

    • Table 4. Quantitative analysis of image enhancement results by different algorithms

      View table

      Table 4. Quantitative analysis of image enhancement results by different algorithms

      AlgorithmImgMSEPSNRSSIMENTR
      12024.0815.070.7611.74
      UDCP21619.1516.040.7510.05
      32495.6114.160.8012.48
      4925.8118.470.807.21
      18106.139.0430.4911.60
      MIP2990.1018.170.8110.71
      34071.1612.030.6612.59
      44331.9611.770.234.72
      1571.4120.560.8512.48
      Ref. [12]21514.7116.330.5111.15
      35318.6910.870.4510.84
      41228.6417.240.767.60
      11163.1917.470.8012.65
      Ref. [13]21399.4316.670.5910.06
      31077.9417.800.8313.00
      43710.2412.440.7512.51
      1523.3420.940.8512.60
      Proposedalgorithm2856.8118.800.8612.32
      3507.9721.070.8812.91
      4848.6618.840.7512.46

    • Table 5. Comparison of complexity of different algorithms unit: s

      View table

      Table 5. Comparison of complexity of different algorithms unit: s

      Algorithm200 pixel×300 pixel300 pixel×400 pixel400 pixel×600 pixel600 pixel×800 pixel800 pixel×1200 pixel1200 pixel×1800 pixel
      UDCP1.7023.4687.02914.20029.21365.533
      MIP1.4532.9686.12412.29425.05756.831
      Ref. [12]3.9377.95116.27832.82166.611150.840
      Ref. [13]0.1590.2880.6591.3012.4435.970
      Proposed algorithm0.0380.0810.2060.4230.8211.862
      Gain4.1803.5503.2003.0802.9803.210
    Tools

    Get Citation

    Copy Citation Text

    Wei Guo, Youbo Zhang, Yue Zhou, Gaofei Xu, Guangwei Li. Rapid Deep-Sea Image Restoration Algorithm Applied to Unmanned Underwater Vehicles[J]. Acta Optica Sinica, 2022, 42(4): 0410002

    Download Citation

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

    Category: Image Processing

    Received: Apr. 30, 2021

    Accepted: Aug. 20, 2021

    Published Online: Jan. 29, 2022

    The Author Email: Zhang Youbo (youbozhang616@163.com)

    DOI:10.3788/AOS202242.0410002

    Topics

    laser devices and laser physics
    Lasers and Laser Optics
    Laser physics
    laser manufacturing
    Instrumentation, Measurement and Metrology