Detection and Analysis of Anomalous Permanent Scatterers in TerraSAR-X Data

Kurzfassung

Permanent Scatterer Interferometry (PSInSAR) leads to the formation of networks of targets with a stable behavior in time, used to overcome decorrelation phenomena induced by long revisiting periods of the same area. This technique has been mainly used for the computation of deformation time series in the form of displacement maps. However, the results depend on the quality of the permanent scatterers (PS) extraction methods and on each individual data set. We select the PS candidates from a set of co-registered SLC products using the spectral behavior of stable points, that significantly differs from the one exhibited by distributed targets. We perform a classification of the parameters used for the extraction process, such as the number of acquisitions and decision thresholds (window sizes, SNR threshold etc.). Apart from analyzing the implication of traditional error sources, such as APS and baseline values, PS with anomalous behavior will be identified, isolated and monitored for the duration of the different time series. Causes for the apparition of false detected targets will be studied, based on the amplitude analysis and an in-depth study of each individual scenery. In order to ensure scenario diversity for the validation of the results, three data sets are being analyzed. The data consist of TerraSAR-X StripMap images taken over urban regions from Bucharest (32 images), Valencia (8 images) and San Francisco (7 images) in the frame of MTH1628, MTH0302 and MTH1118 proposals.