Improving InSAR-based ground deformation monitoring using an RBF-enhanced SBAS method
IEEE Geoscience and Remote Sensing Letters, vol.23, 2026 (SCI-Expanded, Scopus)
- Publication Type: Article / Article
- Volume: 23
- Publication Date: 2026
- Doi Number: 10.1109/lgrs.2026.3709271
- Journal Name: IEEE Geoscience and Remote Sensing Letters
- Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Aerospace Database, Compendex, Geobase, INSPEC, Materials Science & Engineering Collection (ProQuest), Technology Collection (ProQuest)
- Keywords: Interferometric synthetic-aperture radar (InSAR), radial basis functions (RBFs), small baseline subset (SBAS), spatial correlation, time-series deformation analysis
- Ataturk University Affiliated: No
Abstract
The Small Baseline Subset (SBAS) method is widely used to retrieve deformation time series over distributed scatterers, but conventional pixel-wise SBAS (PW-SBAS) ignores the spatial correlation of interferometric phase. This often reduces spatial coverage and increases noise, especially in temporally decorrelated areas such as vegetation and water-bounded regions. To overcome this limitation, we propose RBF-SBAS, which integrates radial basis functions (RBFs) into the SBAS framework to model spatial relationships between neighboring pixels. By limiting each RBF to a local neighborhood, the method improves spatial continuity and inversion stability while enabling a sparse formulation that reduces memory and runtime. Tests using Sentinel-1 data from two study areas, validated with independent GNSS measurements, show that RBF-SBAS significantly increases the number of successfully inverted pixels while achieving comparable or better accuracy than PW-SBAS.