• Journal of Atmospheric and Environmental Optics
  • Vol. 18, Issue 4, 371 (2023)
ZHANG Sugui1,2, ZHANG Jingjing1,2,*, XUN Lina1,2, SUN Xiaobing3..., XIONG Wei3, YAN Qing1,2 and LI Sui4|Show fewer author(s)
Author Affiliations
  • 1Key Laboratory of Intelligent Computing and Signal Processing of Ministry of Education,Anhui University, Hefei 230601, China
  • 2School of Electrical Engineering and Automation, Anhui University, Hefei 230601, China
  • 3Key Laboratory of Optical Calibration and Characterization, Chinese Academy of Sciences, Hefei 230031, China
  • 4Anhui Wenda University of Information Engineering, Hefei 231201, China
  • show less
    DOI: 10.3969/j.issn.1673-6141.2023.04.009 Cite this Article
    Sugui ZHANG, Jingjing ZHANG, Lina XUN, Xiaobing SUN, Wei XIONG, Qing YAN, Sui LI. Cloud detection of GF-5 remote sensing image based on multimodal fusion[J]. Journal of Atmospheric and Environmental Optics, 2023, 18(4): 371 Copy Citation Text show less
    The overall network model architecture
    Fig. 1. The overall network model architecture
    Channel-spatial attention module
    Fig. 2. Channel-spatial attention module
    The module of iAFF
    Fig. 3. The module of iAFF
    MS-CAM
    Fig. 4. MS-CAM
    Comparison of visualization experimental results of different methods
    Fig. 5. Comparison of visualization experimental results of different methods
    Visualization of cloud detection results. (a) Remote sensing imagery of thin cloud areas; (b) ground truth; (c) the results using reflectance information as input; (d) the results using reflectance and polarization information
    Fig. 6. Visualization of cloud detection results. (a) Remote sensing imagery of thin cloud areas; (b) ground truth; (c) the results using reflectance information as input; (d) the results using reflectance and polarization information
    波段/nm波段宽度/nm偏振与否
    443443~453
    490480~500
    565555~575
    670660~680
    763758~768
    765745~785
    865845~885
    910900~920
    Table 1. Characteristics of DPC data bands
    实验方法PA/%PCA/%Rcall/%MIoU/%
    FCN88.6484.5388.5179.16
    SegNet86.2380.5890.9175.74
    DeepLab v191.8988.7091.7884.62
    U-Net92.1394.1093.6784.31
    DANet90.1286.2290.1081.58
    BiSeNet91.2988.0490.9383.55
    This work93.9195.5494.9987.62
    Table 2. Cloud detection accuracy of different methods
    实验组别实验方法PA/%PCA/%Rcall/%MIoU/%
    (a)3DU-Net + R90.8490.6194.8982.26
    (b)3DU-Net + R+ P92.1394.1093.6784.31
    (c)3DU-Net + CSAM + R92.5691.5791.7086.01
    (d)3DU-Net + CSAM + R + P93.0992.0592.0386.33
    (e)3DU-Net + CSAM + iAFF + R93.3392.5292.4887.36
    (f)3DU-Net + CSAM + iAFF + R + P93.9195.5494.9987.62
    Table 3. Cloud detection results of ablation experiments
    Sugui ZHANG, Jingjing ZHANG, Lina XUN, Xiaobing SUN, Wei XIONG, Qing YAN, Sui LI. Cloud detection of GF-5 remote sensing image based on multimodal fusion[J]. Journal of Atmospheric and Environmental Optics, 2023, 18(4): 371
    Download Citation