• Opto-Electronic Engineering
  • Vol. 49, Issue 6, 210423 (2022)
Qian Cheng1、2, Xinjian Gao1、*, Jun Gao1、2, Xin Wang1、2, Tianyi Dang1、2, and Yuan Yan1、2
Author Affiliations
  • 1School of Computer and Information, Hefei University of Technology, Hefei, Anhui 230009, China
  • 2Image Information Processing Laboratory, Hefei University of Technology, Hefei, Anhui 230009, China
  • show less
    DOI: 10.12086/oee.2022.210423 Cite this Article
    Qian Cheng, Xinjian Gao, Jun Gao, Xin Wang, Tianyi Dang, Yuan Yan. A generative method for atmospheric polarization modelling based on neighborhood constraint[J]. Opto-Electronic Engineering, 2022, 49(6): 210423 Copy Citation Text show less
    Polarization angle distribution of the atmospheric polarization under different occlusions
    Fig. 1. Polarization angle distribution of the atmospheric polarization under different occlusions
    Network structure of the atmospheric polarization mode generation based on neighborhood constraints
    Fig. 2. Network structure of the atmospheric polarization mode generation based on neighborhood constraints
    Schematic diagram of the binary mask update for the area recognition
    Fig. 3. Schematic diagram of the binary mask update for the area recognition
    Atmospheric polarization mode masks
    Fig. 4. Atmospheric polarization mode masks
    Angular error results of the solar meridian under different masking types
    Fig. 5. Angular error results of the solar meridian under different masking types
    Reconstruction result of measured atmospheric polarization mode
    Fig. 6. Reconstruction result of measured atmospheric polarization mode
    Comparison of the solar meridian angle error before and after reconstruction
    Fig. 7. Comparison of the solar meridian angle error before and after reconstruction
    Comparison of the results of different reconstruction methods
    Fig. 8. Comparison of the results of different reconstruction methods
    Constrained ablation results of the solar meridian
    Fig. 9. Constrained ablation results of the solar meridian
    Influence of different values ofλ2 on the performance of this method
    Fig. 10. Influence of different values of λ2 on the performance of this method
    MethodMask type
    ABCD
    SSIMContext Enconder PConv PRVS NCAPG without NFR NCAPG 0.858 0.881 0.864 0.860 0.879 0.853 0.873 0.864 0.856 0.872 0.816 0.823 0.826 0.819 0.8440.819 0.827 0.832 0.820 0.851
    PSNRContext Enconder PConv PRVS NCAPG without NFR NCAPG 25.87 26.23 26.85 25.91 27.3624.48 26.56 25.79 24.53 26.7221.59 21.76 22.51 21.54 23.1322.21 22.74 23.55 22.13 24.48
    MSEContext Enconder PConv PRVS NCAPG without NFR NCAPG 0.045 0.038 0.036 0.043 0.0340.067 0.064 0.057 0.066 0.0530.088 0.085 0.080 0.087 0.0730.079 0.076 0.071 0.081 0.062
    Table 1. Quantitative analysis of the results of different reconstruction methods
    Method Coverage/%
    1020304050
    PConv3.23°3.51°4.26°5.13°6.95°
    PRVS3.02°3.36°3.94°4.97°6.23°
    DeepFillv22.84°3.15°3.79°4.62°5.96°
    NCAPG without APCC2.61°3.01°3.67°4.45°5.42°
    NCAPG2.56°2.78°3.14°3.51°4.25°
    Table 2. Comparison of navigation angle errors of different reconstruction methods
    Qian Cheng, Xinjian Gao, Jun Gao, Xin Wang, Tianyi Dang, Yuan Yan. A generative method for atmospheric polarization modelling based on neighborhood constraint[J]. Opto-Electronic Engineering, 2022, 49(6): 210423
    Download Citation