Recent advances in multi-temporal Differential Synthetic Aperture Radar (SAR) Interferometry (DInSAR) have greatly improved our capability to monitor geological processes. Ground motion studies using DInSAR require both the availability of good quality input data and rigorous approaches to exploit the retrieved Time Series (TS) at their full potential. In this work we present a methodology for DInSAR TS analysis, with particular focus on landslides and subsidence phenomena. The proposed methodology consists of three main steps: (1) pre-processing, i.e., assessment of a SAR Dataset Quality Index (SDQI) (2) post-processing, i.e., application of empirical/stochastic methods to improve the TS quality, and (3) trend analysis, i.e., comparative implementation of methodologies for automatic TS analysis. Tests were carried out on TS datasets retrieved from processing of SAR imagery acquired by different radar sensors (i.e., ERS-1/2 SAR, RADARSAT-1, ENVISAT ASAR, ALOS PALSAR, TerraSAR-X, COSMO-SkyMed) using advanced DInSAR techniques (i.e., SqueeSARTM, PSInSARTM, SPN and SBAS). The obtained values of SDQI are discussed against the technical parameters of each data stack (e.g., radar band, number of SAR scenes, temporal coverage, revisiting time), the retrieved coverage of the DInSAR results, and the constraints related to the characterization of the investigated geological processes. Empirical and stochastic approaches were used to demonstrate how the quality of the TS can be improved after the SAR processing, and examples are discussed to mitigate phase unwrapping errors, and remove regional trends, noise and anomalies. Performance assessment of recently developed methods of trend analysis (i.e., PS-Time, Deviation Index and velocity TS) was conducted on two selected study areas in Northern Italy affected by land subsidence and landslides. Results show that the automatic detection of motion trends enhances the interpretation of DInSAR data, since it provides an objective picture of the deformation behaviour recorded through TS and therefore contributes to the understanding of the on-going geological processes.
A User-Oriented Methodology for DInSAR Time Series Analysis and Interpretation: Landslides and Subsidence Case Studies
NOTTI, DAVIDE;MEISINA, CLAUDIA;ZUCCA, FRANCESCO
2015-01-01
Abstract
Recent advances in multi-temporal Differential Synthetic Aperture Radar (SAR) Interferometry (DInSAR) have greatly improved our capability to monitor geological processes. Ground motion studies using DInSAR require both the availability of good quality input data and rigorous approaches to exploit the retrieved Time Series (TS) at their full potential. In this work we present a methodology for DInSAR TS analysis, with particular focus on landslides and subsidence phenomena. The proposed methodology consists of three main steps: (1) pre-processing, i.e., assessment of a SAR Dataset Quality Index (SDQI) (2) post-processing, i.e., application of empirical/stochastic methods to improve the TS quality, and (3) trend analysis, i.e., comparative implementation of methodologies for automatic TS analysis. Tests were carried out on TS datasets retrieved from processing of SAR imagery acquired by different radar sensors (i.e., ERS-1/2 SAR, RADARSAT-1, ENVISAT ASAR, ALOS PALSAR, TerraSAR-X, COSMO-SkyMed) using advanced DInSAR techniques (i.e., SqueeSARTM, PSInSARTM, SPN and SBAS). The obtained values of SDQI are discussed against the technical parameters of each data stack (e.g., radar band, number of SAR scenes, temporal coverage, revisiting time), the retrieved coverage of the DInSAR results, and the constraints related to the characterization of the investigated geological processes. Empirical and stochastic approaches were used to demonstrate how the quality of the TS can be improved after the SAR processing, and examples are discussed to mitigate phase unwrapping errors, and remove regional trends, noise and anomalies. Performance assessment of recently developed methods of trend analysis (i.e., PS-Time, Deviation Index and velocity TS) was conducted on two selected study areas in Northern Italy affected by land subsidence and landslides. Results show that the automatic detection of motion trends enhances the interpretation of DInSAR data, since it provides an objective picture of the deformation behaviour recorded through TS and therefore contributes to the understanding of the on-going geological processes.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
