TIME-SERIES LANDSLIDE MONITORING BASED ON STAMPS-SBAS: A CASE STUDY IN LUSHAN, TAIWAN

Multiple temporal InSAR (MTInSAR) is a useful tool that has been widely used in slow-moving landslides detection and monitoring. However, the serious decorrelation can affect the scatterer selection and phase unwrapping during the procedures of MTInSAR when the amount of SAR images is small. In this study, a modified StaMPS-SBAS is provided to time-series deformation monitoring for landslides detection and monitoring. Multilook and nonlocal filtering operations are adopted to improve the coherence and phase quality before StaMPS-SBAS. The time-series deformation is validated with three collected GPS stations. The results found that: (1) Several landslides with average rate over 40mm/yr in the research of interest are observed. (2) The reginal maximum rate reached to - 92.7mm/yr. (3) The RMSE of time-series deformation in the selected GPS station is 4.4 mm.