A clustering approach for studying ground deformation trends in Campania Region through PS-InSAR™ time series analysis

Recent development of remote sensing techniques brings new possibilities of precise ground movement measurements. One of these possibilities is the PS-InSARTM method. It allows to detect small and long period ground deformations on large areas. In particular, the PS-InSAR technique involves interferometric phase comparison of several radar images of the same scene taken at different times along the same orbit by satellite radar sensors. The ground deformations occurring in the Campania Region has been recently investigated by applying this interferometric technique in a peculiar project developed by Italian Ministry of Environment. The outcome consists in a very large and complex database that cannot be easily analysed with traditional tools. In order to improve the interpretability of this dataset and to support the study of the geologists of this project, we apply a clustering data mining approach on a sample area and identify homogeneous clusters of ground deformation velocity trends. We report in this paper the main results of this analysis, showing how the chosen clustering method allows a coherent geological interpretation of this large and scattered dataset.