• Laser & Optoelectronics Progress
  • Vol. 59, Issue 4, 0410001 (2022)
Xinyuan Gui, Ran Zhang, Haoyuan Cheng, Lianbiao Tian, and Jinkui Chu*
Author Affiliations
  • Key Laboratory for Micro/Nano Technology and System of Liaoning Province, School of Mechanical Engineering, Dalian University of Technology, Dalian , Liaoning 116024, China
  • show less
    DOI: 10.3788/LOP202259.0410001 Cite this Article Set citation alerts
    Xinyuan Gui, Ran Zhang, Haoyuan Cheng, Lianbiao Tian, Jinkui Chu. Multi-Turbidity Underwater Image Restoration Based on Neural Network and Polarization Imaging[J]. Laser & Optoelectronics Progress, 2022, 59(4): 0410001 Copy Citation Text show less
    Mueller matrix image
    Fig. 1. Mueller matrix image
    Experimental schematic of polarizing dataset shooting
    Fig. 2. Experimental schematic of polarizing dataset shooting
    Intensity images took under turbid and clean water
    Fig. 3. Intensity images took under turbid and clean water
    Neural network structure
    Fig. 4. Neural network structure
    Sliding-window overlay recovery method
    Fig. 5. Sliding-window overlay recovery method
    Comparison of recovery effects. (a) Step of 8; (b) step of 1
    Fig. 6. Comparison of recovery effects. (a) Step of 8; (b) step of 1
    Comparison of image recovery effects between different methods. (a) Original intensity image under turbid water; (b) multi-band fusion method[17]; (c) neural network method[18]; (d) wavelet defogging method[19]; (e) method proposed in this paper
    Fig. 7. Comparison of image recovery effects between different methods. (a) Original intensity image under turbid water; (b) multi-band fusion method[17]; (c) neural network method[18]; (d) wavelet defogging method[19]; (e) method proposed in this paper
    Image1 mL milk2 mL milk3 mL milk4 mL milk
    Turbid image10.6410.109.698.97
    Recovered images with step of 814.7513.7012.9611.70
    Recovered images with step of 415.2113.9813.5512.60
    Recovered images with step of 215.6414.1113.8913.06
    Recovered images with step of 8115.9014.3214.1613.42
    Table 1. PSNR comparison of recovered images
    Method1 mL milk2 mL milk3 mL milk4 mL milk
    Method in Ref.[170.270.290.160.03
    Method in Ref.[180.390.410.370.27
    Method in Ref.[190.220.180.180.15
    Ours0.550.490.440.41
    Table 2. SSIM of images restored by different methods
    Method1 mL milk2 mL milk3 mL milk4 mL milk
    Method in Ref.[170.600.440.380.35
    Method in Ref.[182.341.671.341.25
    Method in Ref.[191.571.341.201.05
    Ours1.941.871.731.72
    Table 3. EME of images restored by different methods
    Xinyuan Gui, Ran Zhang, Haoyuan Cheng, Lianbiao Tian, Jinkui Chu. Multi-Turbidity Underwater Image Restoration Based on Neural Network and Polarization Imaging[J]. Laser & Optoelectronics Progress, 2022, 59(4): 0410001
    Download Citation