Laser & Optoelectronics Progress, Volume. 61, Issue 4, 0428013(2024)

Deep Learning Cloud Detection Based on Regression Analysis of Temporal Data

Yanan Tian, Yunling Li*, Lin Sun**, Shulin Pang, and Ping Zhang
Author Affiliations
  • College of Geodesy and Geomatics, Shandong University of Science and Technology, Qingdao 266590, Shandong, China
  • show less
    Figures & Tables(8)
    Apparent reflectance analysis of time series data, the dash shows the reflectivity line between cloudless image and thick cloud image in December
    Slope band analysis
    UNet++ model structure
    Model training parameters
    Comparison of cloud identification results between short-time-series dataset and single-temporal dataset. (a) Over the barren; (b) over the city; (c) over the forest; (d) over the grassland; (e) over the shrub; (f) over the ice and snow; (g) over water; (h) over wetland
    Accuracy comparison of models. (a) Over the barren; (b) over the city; (c) over the forest; (d) over the grassland; (e) over the shrub; (f) over the ice and snow; (g) over water; (h) over wetland
    Impact of threshold on precision and recall
    • Table 1. Assessment of accuracy

      View table

      Table 1. Assessment of accuracy

      DatasetModelαOA /%αF1_score /%αFA /%αMAR /%
      Single-temporal+BandsU-Net95.7293.174.344.11
      SegNet94.1290.685.786.08
      UNet++95.5592.914.554.23
      Multi-temporal Slope+BandsSlope-UNet++98.2196.281.073.12
    Tools

    Get Citation

    Copy Citation Text

    Yanan Tian, Yunling Li, Lin Sun, Shulin Pang, Ping Zhang. Deep Learning Cloud Detection Based on Regression Analysis of Temporal Data[J]. Laser & Optoelectronics Progress, 2024, 61(4): 0428013

    Download Citation

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

    Category: Remote Sensing and Sensors

    Received: Dec. 26, 2022

    Accepted: Mar. 15, 2023

    Published Online: Feb. 26, 2024

    The Author Email: Li Yunling (742984421@qq.com), Sun Lin (sunlin6@126.com)

    DOI:10.3788/LOP223399

    Topics