Remote Sensing and Digital Image Processing

Soon after the ?rst attempts at DelftUniversityof Technology to apply the radar interferometric technique for the monitoring of subsidence due to gas extractionintheprovinceofGroningen, theNetherlands, itwasrecognizedby Usai and Hanssen (1997) that man-made features remained coherent in radar interferograms over long time spans, while their surrounding was completely decorrelated. This particular area in the northern part of the Netherlands is well-knownforitssubsidence. Duetotheslowsubsidencerate themaximum is approximately 1 cm/y long temporal baselines needed to be used. Even thoughonlyinterferogramswithshortperpendicularbaselinesweregenerated, temporal decorrelation severely limited the analysis, see (Usai, 1997, 2000; Usai and Klees, 1999). The Groningen data set was also used by Hanssen (1998), who analyzed artifacts of atmospheric origin in coherent interfe- grams with short temporal baselines. Aside from temporal and geometrical decorrelation, atmospheric signal is the main problem for the interpretation ofinterferometricsignal of current day spaceborne sensors on board, e. g., ERS, ENVISAT and RADARSAT (Hanssen, 2001). ThePermanentScatterers(PS)Techniquewasdevelopedshortlyafter, see (Ferrettietal.,2000a,2001). Itaimstobypasstheproblemofgeometricaland temporal decorrelation by considering time-coherent pixels. Furthermore, by using a large amount of data, atmospheric signal is estimated and corrected for. The PS technique o?ers a convenient processing framework that enables the use of all acquired images, irrespective of baseline, and a parameter estimation strategy for interferograms with low spatial coherence. The - vantages of this method can be measured from the increasing attention it has received at major conferences For example, in the proceedings of the IGARSS conferences of 1999 to 2003 there are respectively 1, 5, 4, 17 and 26 directreferencestothetermPermanentScatterer."