Learning U-Net without forgetting for near real-time wildfire monitoring by the fusion of SAR and optical time series

被引:53
作者
Zhang, Puzhao [1 ]
Ban, Yifang [1 ]
Nascetti, Andrea [1 ]
机构
[1] KTH Royal Inst Technol, Div Geoinformat, Teknikringen 10A, S-11428 Stockholm, Sweden
关键词
Wildfire; Burned area; Change detection; Deep learning; U-Net; SAR; Sentinel-1; Sentinel-2; NORMALIZED BURN RATIO; FIRE SEVERITY; LANDSAT; CLASSIFICATION; CALIBRATION; INTEGRATION; IMAGES; RDNBR; SCARS; INDEX;
D O I
10.1016/j.rse.2021.112467
中图分类号
X [环境科学、安全科学];
学科分类号
08 ; 0830 ;
摘要
Wildfires are increasing in intensity and frequency across the globe due to climate change and rising global temperature. Development of novel approach to Monitor wildfire progressions in near real-time is therefore of critical importance for emergency responses. The objective of this research is to investigate continuous learning with U-Net by exploiting both Sentinel-1 SAR and Sentinel-2 MSI time series for increasing the frequency and accuracy of wildfire progression mapping. In this study, optical-based burned areas prior to each SAR acquisition (when available) were accumulated into SAR-based pseudo progression masks to train a deep residual U-Net model. Unlike multi-temporal fusion of SAR and optical data, the temporal fusion of progression masks allows us to track as many wildfire progressions as possible. Specifically, two approaches were investigated to train the deep residual U-Net model for continuous learning: 1) Continuous joint training (CJT) with all historical data (including both SAR and optical data); 2) Learning without forgetting (LwF) based on newly incoming data alone (SAR or optical). For LwF, a mean squared loss was integrated to keep the capabilities learned before and prevent it from overfitting to newly incoming data only. By fusing optical-based burned areas, SAR-based progression pseudo masks improve significantly, which benefits both data sampling and model training, considering the challenges in SAR-based change extraction attributed to the variability in SAR backscatter of the surrounding environments. Pre-trained ResNet was frozen as the encoder of the U-Net model, and the decoder part was trained to further refine the derived burned area maps in a progression-wise manner. The experimental results demonstrated that LwF has the potential to match CJT in terms of the agreement between SAR-based results and optical-based ground truth, achieving a F1 score of 0.8423 on the Sydney Fire (2019-2020) and 0.7807 on the Chuckegg Creek Fire (2019). We also observed that the SAR cross-polarization ratio (VH/VV) shows good capability in suppressing multiplicative noise and detecting burned areas when VH and VV have diverse temporal behaviors.
引用
收藏
页数:12
相关论文
共 51 条
  • [1] TerraSAR-X Data for Burn Severity Evaluation in Mediterranean Forests on Sloped Terrain
    Andrei Tanase, Mihai
    Perez-Cabello, Fernando
    de la Riva, Juan
    Santoro, Maurizio
    [J]. IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2010, 48 (02): : 917 - 929
  • [2] Deep semantic segmentation of natural and medical images: a review
    Asgari Taghanaki, Saeid
    Abhishek, Kumar
    Cohen, Joseph Paul
    Cohen-Adad, Julien
    Hamarneh, Ghassan
    [J]. ARTIFICIAL INTELLIGENCE REVIEW, 2021, 54 (01) : 137 - 178
  • [3] Synergy of multitemporal ERS-1 SAR and Landsat TM data for classification of agricultural crops
    Ban, YF
    [J]. CANADIAN JOURNAL OF REMOTE SENSING, 2003, 29 (04) : 518 - 526
  • [4] Near Real-Time Wildfire Progression Monitoring with Sentinel-1 SAR Time Series and Deep Learning
    Ban, Yifang
    Zhang, Puzhao
    Nascetti, Andrea
    Bevington, Alexandre R.
    Wulder, Michael A.
    [J]. SCIENTIFIC REPORTS, 2020, 10 (01)
  • [5] Ban YF, 2016, REMOTE SENS DIGIT IM, V20, P19, DOI 10.1007/978-3-319-47037-5_2
  • [6] Object-Based Fusion of Multitemporal Multiangle ENVISAT ASAR and HJ-1B Multispectral Data for Urban Land-Cover Mapping
    Ban, Yifang
    Jacob, Alexander
    [J]. IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2013, 51 (04): : 1998 - 2006
  • [7] Fusion of Quickbird MS and RADARSAT SAR data for urban land-cover mapping: object-based and knowledge-based approach
    Ban, Yifang
    Hu, Hongtao
    Rangel, I. M.
    [J]. INTERNATIONAL JOURNAL OF REMOTE SENSING, 2010, 31 (06) : 1391 - 1410
  • [8] Evaluation of backscatter coefficient temporal indices for burned area mapping
    Belenguer-Plomer, Miguel A.
    Chuvieco, Emilio
    Tanase, Mihai A.
    [J]. ACTIVE AND PASSIVE MICROWAVE REMOTE SENSING FOR ENVIRONMENTAL MONITORING III, 2019, 11154
  • [9] Burned area detection and mapping using Sentinel-1 backscatter coefficient and thermal anomalies
    Belenguer-Plomer, Miguel A.
    Tanase, Mihai A.
    Fernandez-Carrillo, Angel
    Chuvieco, Emilio
    [J]. REMOTE SENSING OF ENVIRONMENT, 2019, 233
  • [10] Remote monitoring of spatial and temporal surface soil moisture in fire disturbed boreal forest ecosystems with ERS SAR imagery
    Bourgeau-Chavez, L. L.
    Kasischke, E. S.
    Riordan, K.
    Brunzell, S.
    Nolan, M.
    Hyer, E.
    Slawski, J.
    Medvecz, M.
    Walters, T.
    Ames, S.
    [J]. INTERNATIONAL JOURNAL OF REMOTE SENSING, 2007, 28 (10) : 2133 - 2162