Ground Movement Classification Using Statistical Tests Over Persistent Scatterer Interferometry Time Series

Abstract

This study proposes modifications to an existing automatic classification method of Persistent Scatterers Interferometry (PSI) time series (TS) and a new procedure to classify ground movements into seven classes. We also represent a technique to detect TSs affected by phase unwrapping errors and a reclassification part to detect stable points, which are incorrectly classified as moving points using the original method. Around 60 km2 of Catalunya were classified using Sentinel-1 images and a PSI technique. The proposed method classified 78359 PS TS. This study provided the spatial distribution of ground movement classes and detected several time series anomalies.