Analysis of Highly Non-linear Deformations Due to Mining Activity with Dinsar: Psic4 Test Site

DInSAR techniques with multiple images are aimed to extract the deformation series from a selected scene using a set of interferograms acquired in different temporal instants. The detection of highly non-linear deformation patterns is a demanding issue for these techniques. In order to get a high quality and reduced set of interferograms, a new selection method is proposed based on a Minimum Spanning Tree algorithm. Furthermore, two developments on the multiple image DInSAR algorithm Coherent Pixels Technique (CPT) are presented in order to improve non-linear component extraction: a multilayer processing which iteratively adds quality layers of selected points and a temporal block processing. Results on the Gardanne test site (France) are presented.