ESA Earth Home Missions Data Products Resources Applications
EO Data Access
How to Apply
How to Access
Site Map
Frequently asked questions
Terms of use
Contact us



A Simple Robust Two-Scale Phase Component Inversion Scheme for PSInSAR

Bernhard Rabus(1) and Parwant Ghuman(1)

(1) MDA, 1300 Commerce Parkway, Richmond, V7S 1L5, Canada


We have developed a novel method of persistent point scatterer (PS) phase component inversion that has proven to yield significantly better results for smaller interfeormetric stacks and non-linear motion histories compared to the traditional iterative linear phase plane modeling which is still frequently used by commercial and scientific PS InSAR solutions. Different from the latter our new method is more robust and allows full automation. Key features of the new approach are: (1) A new automated way of selecting the optimum global reference point. (2) A split of the phase signal in two spatial scales: This is carried out by a-priori creation of a 'pseudo-atmospheric stack' using a strong spatial filter on scales about five times the atmospheric correlation length (~15 km). It is assumed that the external DEM used for initial flattening is unbiased at this scale and that the temporal distribution of the SAR acquisitions is quasi-uniform. This allows separation of true atmospheric phases from surface displacement contributions at the large spatial scale via temporal low-pass filtering. (3) Automated estimation of the optimum height error is carried out for the small spatial scale component of the phase signal without assuming a displacement model. The only assumption made is a piece- wise 'smoothness' of the displacement time series. Result of this separation step is the non-linear displacement at small spatial scale. (4) Byproducts to the method are a number of quality measures that allow a robust a-posteriori selection of reliable points. Additionally customer provided constraints on the displacement time series can be assimilated at this step. (5) The simple automated design is geared for fast throughput of large stacks consisting of entire scenes. Usually, a single iteration is sufficient to achieve reliable results on the mm-accuarcy level for empirically well-balanced stacks with 20 scenes or more.

In the presentation we will present our method and illustrate its performance using different data sets from the RADARSAT-1 sensor.


Keywords: ESA European Space Agency - Agence spatiale europeenne, observation de la terre, earth observation, satellite remote sensing, teledetection, geophysique, altimetrie, radar, chimique atmospherique, geophysics, altimetry, radar, atmospheric chemistry