INSAR DEFORMATION TIME SERIES CLASSIFICATION USING A CONVOLUTIONAL NEURAL NETWORK

Abstract. Temporal analysis of deformations Time Series (TS) provides detailed information various natural and humanmade displacements. Interferometric Synthetic Aperture Radar (InSAR) generates millimetre-scale products, indicating the chronicle behaviour detected targets via TS products. Deep Learning (DL) can handle a massive load InSAR to categorize significant movements from non-moving targets. To this end, we employed supervised Convolutional Neural Network (CNN) model distinguish five trends, including Stable, Linear, Quadratic, Bilinear, Phase Unwrapping Error (PUE). Considering several arguments in CNN model, trained numerous combinations explore most accurate combination 5000 samples extracted Persistent Scatterer Interferometry (PSI) technique Sentinel-1 images over Granada region, Spain. The overall accuracy exceeds 92%. Deformations three cases landslides were also same area, Cortijo de Lorenzo, El Arrecife, Rules Viaduct areas.