Opto-Electronic Engineering, Volume. 50, Issue 2, 220207(2023)

Multi-frame blind deconvolution of solar images via second-order total generalized variation

Shuai Wang1...2, Chunyuan He1,2, Huiqin Rong1,2, Hua Bao3,4,*, Jialin Hou5, and Changhui Rao34 |Show fewer author(s)
Author Affiliations
  • 1Yangtze Delta Region Institute (Quzhou), University of Electronic Science and Technology of China, Quzhou, Zhejiang 324003, China
  • 2University of Electronic Science and Technology of China, Chengdu, Sichuan 611731, China
  • 3Key Laboratory on Adaptive Optics, Chinese Academy of Sciences, Chengdu, Sichuan 610209, China
  • 4Institute of Optics and Electronics, Chinese Academy of Sciences, Chengdu, Sichuan 610209, China
  • 566389 Troops, Zhengzhou, Henan 450000, China
  • show less
    Figures & Tables(9)
    Flow chart of the solar image reconstruction
    Reconstruction results of the wide field-of-view image. (a) The image with the best overall quality in the input sequence; (b) TVBD; (c) S-TGV; (d) OBD; (e) M-TGV; (f) Speckle reconstruction
    Subregion reconstruction results of the wide field-of-view image. (a) The image with the best overall quality in the input sequence; (b) TVBD; (c) S-TGV; (d) OBD; (e) M-TGV; (f) Speckle reconstruction
    Reconstruction results of different input frames. (a) One frame in the input sequence; (b) ~ (g) The reconstruction results corresponding to input 1, 2, 3, 5, 10 and 20 frames respectively (h) Speckle reconstruction
    The effectiveness of image regularization. (a) One frame in the input sequence; (b) Without regularization; (c) M-TGV; (d) Speckle reconstruction
    Reconstructed images and estimated PSF with different K and β1 values. (a) K=1, β1=0; (b) K=1, β1=10; (c) K=5, β1=0; (d) K=5, β1=10
    Iteration curves of the objective function. (a) Fourth scale; (b) Third scale; (c) Second scale; (d) First scale
    • Table 1. Quantitative evaluation of reconstruction results (PSNR(dB)/SSIM)

      View table
      View in Article

      Table 1. Quantitative evaluation of reconstruction results (PSNR(dB)/SSIM)

      输入序列中总体质量最好帧TVBDS-TGVOBDM-TGV
      米粒132.68/0.906032.05/0.905832.27/0.912732.70/0.908232.79/0.9173
      米粒233.63/0.919833.60/0.926933.45/0.930234.00/0.921934.48/0.9396
      半影29.41/0.843429.00/0.842429.57/0.857629.70/0.854230.38/0.8785
      本影29.65/0.877129.68/0.872130.08/0.892730.09/0.884330.30/0.8975
      整帧图像31.06/0.880630.77/0.879330.96/0.889231.25/0.885531.31/0.8969
    • Table 2. Running time of different algorithms

      View table
      View in Article

      Table 2. Running time of different algorithms

      算法TVBDS-TGVOBDM-TGV
      时间/s715.3712.35156.6127.68
    Tools

    Get Citation

    Copy Citation Text

    Shuai Wang, Chunyuan He, Huiqin Rong, Hua Bao, Jialin Hou, Changhui Rao. Multi-frame blind deconvolution of solar images via second-order total generalized variation[J]. Opto-Electronic Engineering, 2023, 50(2): 220207

    Download Citation

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

    Category: Article

    Received: Aug. 27, 2022

    Accepted: Jan. 3, 2023

    Published Online: Apr. 13, 2023

    The Author Email: Bao Hua (hbao@ioe.ac.cn)

    DOI:10.12086/oee.2023.220207

    Topics