Spatiotemporal mining of ENVISAT SAR interferogram time series over the Haiyuan fault in China

In this paper, an original approach for analyzing InSAR time series is presented. The interferograms forming such time series allow ground deformation occurring between acquisition dates to be measured with high precision. Nevertheless, they can be affected by variations in atmospheric conditions. The proposed approach is designed to handle these varying atmospheric conditions. The stratified atmosphere is first removed and the phase evolution is built using a Small BAseline Subsets (SBAS) strategy. Then, frequent grouped sequential patterns are extracted. These patterns allow InSAR time series to be described spatially and temporally while discarding atmospheric perturbations. Experimental results on an ENVISAT InSAR time series covering the Haiyuan fault in the northeastern boundary of the Tibetan plateau are presented.