Generative networks for spatio-temporal gap filling of Sentinel-2 reflectances

被引:0
|
作者
Gonzalez-Calabuig, Maria [1 ]
Fernandez-Torres, Miguel-angel [1 ]
Camps-Valls, Gustau [1 ]
机构
[1] Univ Valencia, Image Proc Lab, C Cat Agustin Escardino Benlloch,9, Paterna 46980, Valencia, Spain
基金
欧洲研究理事会; 欧盟地平线“2020”;
关键词
Gap-filling; Visible and near-infrared reflectances; Cloud screening; Interpolation; Generative Adversarial Network (GAN); Sentinel-2; REMOTE-SENSING-IMAGE; CLOUD REMOVAL; OPTICAL-DATA; LARGE-SCALE; SAR;
D O I
10.1016/j.isprsjprs.2025.01.016
中图分类号
P9 [自然地理学];
学科分类号
0705 ; 070501 ;
摘要
Earth observation from satellite sensors offers the possibility to monitor natural ecosystems by deriving spatially explicit and temporally resolved biogeophysical parameters. Optical remote sensing, however, suffers from missing data mainly due to the presence of clouds, sensor malfunctioning, and atmospheric conditions. This study proposes a novel deep learning architecture to address gap filling of satellite reflectances, more precisely the visible and near-infrared bands, and illustrates its performance at high-resolution Sentinel-2 data. We introduce GANFilling, a generative adversarial network capable of sequence-to-sequence translation, which comprises convolutional long short-term memory layers to effectively exploit complete dependencies in space- time series data. We focus on Europe and evaluate the method's performance quantitatively (through distortion and perceptual metrics) and qualitatively (via visual inspection and visual quality metrics). Quantitatively, our model offers the best trade-off between denoising corrupted data and preserving noise-free information, underscoring the importance of considering multiple metrics jointly when assessing gap filling tasks. Qualitatively, it successfully deals with various noise sources, such as clouds and missing data, constituting a robust solution to multiple scenarios and settings. We also illustrate and quantify the quality of the generated product in the relevant downstream application of vegetation greenness forecasting, where using GANFilling enhances forecasting in approximately 70% of the considered regions in Europe. This research contributes to underlining the utility of deep learning for Earth observation data, which allows for improved spatially and temporally resolved monitoring of the Earth surface.
引用
收藏
页码:637 / 648
页数:12
相关论文
共 50 条
  • [1] Spatio-temporal fusion for daily Sentinel-2 images
    Wang, Qunming
    Atkinson, Peter M.
    REMOTE SENSING OF ENVIRONMENT, 2018, 204 : 31 - 42
  • [2] Spatio-Temporal Generative Adversarial Networks
    Qin, Chao
    Gao, Xiaoguang
    CHINESE JOURNAL OF ELECTRONICS, 2020, 29 (04) : 623 - 631
  • [3] Spatio-Temporal Generative Adversarial Networks
    QIN Chao
    GAO Xiaoguang
    ChineseJournalofElectronics, 2020, 29 (04) : 623 - 631
  • [4] Distributed spatio-temporal generative adversarial networks
    QIN Chao
    GAO Xiaoguang
    JournalofSystemsEngineeringandElectronics, 2020, 31 (03) : 578 - 592
  • [5] Distributed spatio-temporal generative adversarial networks
    Qin Chao
    Gao Xiaoguang
    JOURNAL OF SYSTEMS ENGINEERING AND ELECTRONICS, 2020, 31 (03) : 578 - 592
  • [6] CLOUD AND SHADOW REMOVAL FOR SENTINEL-2 BY PROGRESSIVELY SPATIO-TEMPORAL PATCH GROUP LEARNING
    Zhang, Qiang
    Yuan, Qiangqiang
    Li, Jie
    Shen, Huanfeng
    Zhang, Liangpei
    2019 IEEE INTERNATIONAL GEOSCIENCE AND REMOTE SENSING SYMPOSIUM (IGARSS 2019), 2019, : 775 - 778
  • [7] Spatio-temporal Evolution of Crop Fields in Sentinel-2 Satellite Image Time Series
    Solano-Correa, Yady Tatiana
    Bovolo, Francesca
    Bruzzone, Lorenzo
    Fernandez-Prieto, Diego
    2017 9TH INTERNATIONAL WORKSHOP ON THE ANALYSIS OF MULTITEMPORAL REMOTE SENSING IMAGES (MULTITEMP), 2017,
  • [8] Generative Adversarial Networks for Spatio-temporal Data: A Survey
    Gao, Nan
    Xue, Hao
    Shao, Wei
    Zhao, Sichen
    Qin, Kyle Kai
    Prabowo, Arian
    Rahaman, Mohammad Saiedur
    Salim, Flora D.
    ACM TRANSACTIONS ON INTELLIGENT SYSTEMS AND TECHNOLOGY, 2022, 13 (02)
  • [9] Sentinel-2 for mapping the spatio-temporal development of submerged aquatic vegetation at Lake Starnberg (Germany)
    Fritz, Christine
    Kuhwald, Katja
    Schneider, Thomas
    Geist, Juergen
    Oppelt, Natascha
    JOURNAL OF LIMNOLOGY, 2019, 78 (01) : 71 - 91
  • [10] Spatio-Temporal Classification Framework for Mapping Woody Vegetation from Multi-Temporal Sentinel-2 Imagery
    Kovacevic, Jovan
    Cvijetinovic, Zeljko
    Lakusic, Dmitar
    Kuzmanovic, Nevena
    Sinzar-Sekulic, Jasmina
    Mitrovic, Momir
    Stancic, Nikola
    Brodic, Nenad
    Mihajlovic, Dragan
    REMOTE SENSING, 2020, 12 (17) : 1 - 23