Laser & Optoelectronics Progress, Volume. 62, Issue 16, 1611003(2025)

Simple Optical System Computational Imaging Method for Remote Sensing Camera Based on MsRN-Deblur TransGAN

Xindong Zhao1,2, Chunyu Liu1、*, Chen Wang1, and Jinghui Huang1,2
Author Affiliations
  • 1Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences, Changchun 130033, Jilin , China
  • 2University of Chinese Academy of Sciences, Beijing 100049, China
  • show less
    Figures & Tables(24)
    Roadmap of principles and ideas for simple optical systems in computational imaging mode
    PSF distribution diagram of optical system
    Optical imaging degradation model
    MsRN-Deblur TransGAN deblurring technique flowchart
    MfFTRM
    MDTA module
    RDB
    GRFF network
    MSCAM
    Multi-scale residual module
    MsRN
    Generator network structure
    Global-local dual-scale discriminating network structure
    Simple lens system optimization results. (a) Cook three-piece lens; (b) MTF curve graph with a cutoff frequency of 100 lp/mm; (c) 0° field of view PSF; (d) 3.1° field of view PSF; (e) 4.5° field of view PSF; (f) RMS wavefront error curve for each field of view of central wavelength
    Simple lens system imaging simulation diagrams based on Zemax. (a) Original image of an airport; (b) blurred simulation diagram of an airport; (c) original image of a sports center; (d) blurred simulation diagram of a sports center
    Eight-lens optical system optimization results. (a) Optical system structure; (b) MTF curve graph with a cutoff frequency of 100 lp/mm; (c) 0° field of view PSF; (d) 3.1° field of view PSF; (e) 4.5° field of view PSF; (f) RMS wavefront error curve for each field of view of central wavelength
    Comparison of different network restoration effects. (a) Simple optical system original blurred images; (b) DeblurGAN restoration results; (c) DeblurGAN-v2 restoration results; (d) Restormer restoration results; (e) restoration results of the proposed network; (f) results of imaging through an eight-lens optical system; (g) original clear images
    Restoration effect of electronic map satellite images. (a) Three-lens simple optical system blurred images; (b) restored images; (c) original clear images
    Comparison of generalization experiment restoration effects
    • Table 1. Simple optical system design parameters

      View table

      Table 1. Simple optical system design parameters

      System parameterValue
      Focal length1062.5 mm
      Full field of view
      System F number5.8
      Pupil diameter183 mm
      Work band380‒760 nm
    • Table 2. Results of ablation experiments of MsRN-Deblur TransGAN modules

      View table

      Table 2. Results of ablation experiments of MsRN-Deblur TransGAN modules

      BaselineMDTARDBGRFFFFNMsRN (with GN)MsRN (with BN)PSNR /dBSSIMParams /106FLOPs /109
      28.350.88312.9678.29
      29.230.89716.5681.32
      31.560.91618.2712.51
      30.360.90413.5720.82
      32.510.92116.8830.64
      33.290.92816.2763.65
    • Table 3. Quantitative comparison of image restoration

      View table

      Table 3. Quantitative comparison of image restoration

      TypePSNR /dBSSIM
      Original blurred image21.630.658
      DeblurGAN1828.350.883
      DeblurGAN-v21930.070.897
      Restormer3031.870.916
      Our model33.290.928
      Eight-lens optical system33.560.931
    • Table 4. Quantitative experimental results of electronic map satellite image restoration

      View table

      Table 4. Quantitative experimental results of electronic map satellite image restoration

      TypePSNR /dBSSIM
      Blurred image22.760.683
      Restored image32.590.916
    • Table 5. Quantitative comparison of generalization experiment

      View table

      Table 5. Quantitative comparison of generalization experiment

      TypePSNR /dBSSIM
      Original blurred image24.430.627
      Multi-scale25.010.686
      Fov-GAN1225.260.748
      Unet28.340.792
      RRG-GAN1330.410.865
      DIFAN530.520.884
      Our model31.680.912
    Tools

    Get Citation

    Copy Citation Text

    Xindong Zhao, Chunyu Liu, Chen Wang, Jinghui Huang. Simple Optical System Computational Imaging Method for Remote Sensing Camera Based on MsRN-Deblur TransGAN[J]. Laser & Optoelectronics Progress, 2025, 62(16): 1611003

    Download Citation

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

    Category: Imaging Systems

    Received: Feb. 14, 2025

    Accepted: Mar. 17, 2025

    Published Online: Aug. 6, 2025

    The Author Email: Chunyu Liu (mmliucy@163.com)

    DOI:10.3788/LOP250635

    CSTR:32186.14.LOP250635

    Topics