Recursive Persistent Scatterer Interferometry

Petar Marinkovic(1) , Freek van Leijen(1) , Gini Ketelaar(1) , and Ramon Hanssen(1)

(1) Delft University of Technology, Klyverweg 1, 2629HS Delft, Netherlands

Abstract

Time series InSAR analysis using Persistent Scatterer (PS) techniques is aimed at the joint estimation of topographic and displacement signal from a number of interferometric combinations. Since the estimated values of both parameters are correlated and error signal due to, e.g., atmospheric signal can significantly affect the adjustment, accurate estimation depends on the availability of a large data stack, i.e., more than 20-30 images. A smaller number of images usually results in problems like detecting persistent scatterers, reducing the atmospheric signal, separating topography and displacement, and phase ambiguity estimation.

Unfortunately, there are many areas in the world where the radar data archive is rather limited and techniques as described above are not sufficiently reliable. Moreover, when a new sensor is launched, it will take many repeat orbits before an analysis can be started. In this paper, the problem of parameter estimation in PSI is approached in a recursive manner, where the precision and reliability of the estimated parameters is a function of the number of acquisitions and their geometric and temporal characteristics. Effectively, this results in estimated parameters even if only few images are available. The quality description of these estimated parameters can consequently be used to determine their significance and reliability. This framework sets the basis for recursive data processing strategies, where new acquisitions can be easily added to an existing data stack, significantly reducing computational requirements.

Potentials and limitations of the presented method are investigated and validated on several test sites with an emphasis on the city of Moscow data set. The Moscow set has an observation time of more than twelve years with ERS-1 and ERS-2 sensors. The deformation maps obtained from a large stack of images (60+) are cross-compared to results obtained from the small ones and ones processed in the recursive manner.

 

Full paper

 

  Higher level                 Last modified: 07.10.03