An Advanced Hybrid Mtinsar Based Approach for Deformation Detection in Mining Region

In this paper, a hybrid multi-temporal interferometric synthetic aperture radar (MTInSAR) method has been devised with two most effective strategies - (i) Component extrAc tion and sElection SAR (CAESAR) and (ii) Interferometric Point Target Analysis (IPTA). The efficacy of this novel approach has been tested while estimating deformation leading to tailing dam failure in Brumadinho mines, Brazil, which happened in January 2019, utilising Sentinel-1 descending Interferometric Wide Single-Look Complex (IW SLC) dataset between 2018-19. The hybrid method is effectively able to identify coherent scatterers approximately thirty-two times larger in numbers than those detected by the IPTA and seven times more than those detected by another popular MTInSAR-based method - Looking inside the Continents from Space using Small BAseline Subset (LiCSBAS). Further, using the time-series results, inverse velocity approach has been applied to predict slope failures in the same area. It is observed that the average deviation of the true time of failure from the predicted time of failure is improved from 56 to 16 days, which accounts for 67% improvement. We denote this difference between true and predicted time of failure as ∆ToF.