Optics and Precision Engineering, Volume. 30, Issue 17, 2155(2022)

Super-resolution reconstruction method for space target images based on dense residual block-based GAN

Haizhao JING1, Jianglin SHI2,3、*, Mengzhe QIU1, Yong QI1,4, and Wenxiao ZHU3
Author Affiliations
  • 1Shaanxi Joint Laboratory of Artificial Intelligence, Shaanxi University of Science and Technology, Xi'an7002, China
  • 2Institute of Systems Engineering, Xi'an Jiaotong University, Xi'an710049, China
  • 3Xi'an Satellite Control Center, Xi'an71004, China
  • 4School of Electronic Information and Artificial Intelligence, Shaanxi University of Science and Technology, Xi'an710021, China
  • show less
    Figures & Tables(11)
    Framework of residual in residual dense block GAN generator network
    Framework of SRGAN generator network
    Framework of SRGAN discriminator network
    Difference between standard and relative discriminators
    Simulation data set (partial) of space target AO images
    Comparison of super-resolution results of AO images from Hubble Telescope 1((a) is a low-resolution simulation adaptive optical image whose original input is 256×256, and (b)-(e) are 1 024×1 024 super resolution image network)
    Comparison of super-resolution results of AO images from sat1 Satellite((a) is a low-resolution simulation adaptive optical image whose original input is 256×256, and (b)-(e) are 1 024×1 024 super resolution image network)
    Comparison of super-resolution results of AO images from Hubble Telescope 2((a) is a low-resolution simulation adaptive optical image whose original input is 256×256, and (b)-(e) are 1 024×1 024 super resolution image network)
    • Table 1. Evaluation index of four super resolution methods of Hubble Telescope 1 AO image

      View table
      View in Article

      Table 1. Evaluation index of four super resolution methods of Hubble Telescope 1 AO image

      方 法PSNRSSIM
      Bicubic24.650.744 9
      Nearest25.370.825 1
      BSRNet27.910.867 4
      Ours28.320.910 5
    • Table 2. Evaluation index of four super resolution methods of Sat1 Satellite AO image

      View table
      View in Article

      Table 2. Evaluation index of four super resolution methods of Sat1 Satellite AO image

      方 法PSNRSSIM
      Bicubic25.470.799 1
      Nearest24.780.784 2
      BSRNet27.690.854 6
      Ours29.930.906 9
    • Table 3. Evaluation index of four super resolution methods of Hubble Telescope 2 AO image

      View table
      View in Article

      Table 3. Evaluation index of four super resolution methods of Hubble Telescope 2 AO image

      方 法PSNRSSIM
      Bicubic25.660.782 5
      Nearest24.940.835 1
      BSRNet27.590.899 7
      Ours30.370.934 2
    Tools

    Get Citation

    Copy Citation Text

    Haizhao JING, Jianglin SHI, Mengzhe QIU, Yong QI, Wenxiao ZHU. Super-resolution reconstruction method for space target images based on dense residual block-based GAN[J]. Optics and Precision Engineering, 2022, 30(17): 2155

    Download Citation

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

    Category: Information Sciences

    Received: Jun. 7, 2022

    Accepted: --

    Published Online: Oct. 20, 2022

    The Author Email: Jianglin SHI (shijianglin89@163.com)

    DOI:10.37188/OPE.20223017.2155

    Topics