Time Series Scattering Power Decomposition Using Ensemble Average in Temporal-Spatial Domains: Application to Forest Disturbance Detection

• Published in: IEEE Geoscience and Remote Sensing Letters Vol. 21; pp. 1 - 5
• Main Authors: Sugimoto, Ryu, Natsuaki, Ryo, Nakamura, Ryosuke, Tsutsumi, Chiaki, Yamaguchi, Yoshio
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 2024; Institute of Electrical and Electronics Engineers (IEEE); The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Decomposition; Deforestation; Detection; Dual polarization; European Space Agency; forest disturbance detection; Forest management; Forests; Rainforests; SAR (radar); Satellite constellations; Scattering; scattering power decomposition; Sentinel-1; Spatial discrimination; Spatial resolution; Synthetic aperture radar; Time series; Time series analysis; tropical forests; Vegetation mapping
• ISSN: 1545-598X; 1558-0571; 1558-0571
• DOI: 10.1109/LGRS.2023.3346378

This letter proposes a novel synthetic aperture radar (SAR) time series analysis method based on the scattering power decomposition algorithm with a reasonable ensemble average in both temporal and spatial domains. We reveal that the ensemble average is effective not only in the spatial domain but also in the temporal-spatial domains in the scattering power decomposition. That is, if we extend the ensemble average window in the temporal domain, the proposed method can accurately achieve volume scattering power with a higher spatial resolution than conventional approaches. The precise volume scattering power serves accurate forest monitoring. As an application, we performed forest disturbance detection in the Amazon rainforest using Sentinel-1 time series data. The proposed method detected the disturbances earlier, in less than 2 months, compared to other methods that take about 3 months.