Wavelet Spatio-Temporal Change Detection on Multitemporal SAR Images

• Published in: IEEE journal of selected topics in applied earth observations and remote sensing Vol. 16; pp. 1 - 12
• Main Authors: Fonseca, Rodney V., Negri, Rogerio G., Pinheiro, Aluisio, Atto, Abdourrahmane
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 01-01-2023; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Approximation; Change detection; Coefficients; Computer Science; Correlation; Detection; Feature extraction; Mathematics; Methods; multitemporal images; Radar polarimetry; remote sensing; Screening; simulated images; Speckle; Synthetic aperture radar; Temporal variations; Time series analysis; Wavelet analysis; Wavelet coefficients; wavelets; wavelets
• ISSN: 1939-1404; 2151-1535
• DOI: 10.1109/JSTARS.2023.3268601

We introduce the WECS (Wavelet Energies Correlation Screening), an unsupervised method to detect spatio-temporal changes on multitemporal SAR images. The procedure is based on wavelet approximation for the multitemporal images, wavelet energy apportionment, and ultra-high dimensional correlation screening for the wavelet coefficients. We show WECS's performance on simulated multitemporal image data. We also evaluate the proposed method on a time series of 85 Sentinel-1 images of a forest region at the border of Brazil and French Guiana. Comparisons with well-known change detection methods found in the literature highlight the proposal's superiority in terms of change detection accuracy. Additionally, the introduced method has simple architecture and low computational cost.