Coherent stacking with TerraSAR-X imagery in urban areas

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

12 Scopus citations

Abstract

The German radar satellite TerraSAR-X was launched in June 2007. It is one of the first satellites to continuously provide space-borne high resolution radar imagery with a slant range resolution in the order of 0.6 m x 1.1 m for civil applications. The sensor, the mission design, the orbit concept, and the SAR processor (TMSP) perfectly support interferometric applications. Naturally, DLR's operational interferometric system GENESIS has been adapted to exploit the innovative high resolution data. Algorithm updates have been proven indispensable due to the more complicated spectral characteristic of the data introduced by the spotlight acquisition mode. Also, the high spatial resolution requires that radargrammetric effects, i.e. local parallaxes, be considered in interferometric processing. Several interferometric methods using high resolution TerraSAR-X data will be presented in detail: conventional interferometric three-dimensional mapping of the Earth's surface, assessment of ground and building deformation using differential interferometry (D-InSAR), and the more advanced persistent scatterer interferometry (PSI). Interferometric imaging is particularly interesting for mapping of urban areas, where standard SAR imagery is hard to interpret due to lay-over, shadow, and multiple scattering effects. Multi-baseline techniques, PSI, and SAR tomography allow resolving these geometric ambiguities and aim at a true 3-D localization of objects and their potential motion vectors. The high-resolution and the so far unchallenged geometric accuracy of TerraSAR-X data push open the door to a new quality of geodetic imaging: • Compared to the ERS-type SARs the number of long-time stable (persistent) scatterers increases dramatically from typically 400 per km2 to more than 30,000 per km2, i.e. to several tens per building, allowing the mapping of the 3-D structure of the city and its deformations (i.e. 4-D). • The resolution is high enough to investigate structural deformation of individual buildings (e.g. due to thermal stress) by D-InSAR for the first time. • The short revisit time of 11 days as compared to the 35 day of ERS and ENVISAT favors a fast build-up of interferometric data stacks. • A wide range of incidence angles supports the 4-D imaging capabilities. The most powerful interferometric imaging method on urban areas is the PSI technique. This technology, although well established, needs to be complemented by tomography when ultimate 3-D/4-D imaging quality is asked for. In the extended layover areas scatterers from the ground and from buildings are mapped into a single resolution cell quite often. The goal of SAR tomography is to estimate the 3-D scatterer distribution rather than only an elevation value for every point in 2-D space. Like PSI it is based on coherent stacks of SAR data and uses the diversity of the look angle i.e. the effective baseline to establish a synthetic aperture in the cross-range (elevation) direction. Therefore, it can be considered a natural extension of the PSI technique which was initially developed for the special case of a single scatterer inside the resolution cell. In this presentation we will illustrate the information content of spotlight SAR images in urban areas and demonstrate the technological potential of high resolution SAR interferometry from TerraSAR-X by PSI and tomography processing examples.

Original languageEnglish
Title of host publication2009 Joint Urban Remote Sensing Event
DOIs
StatePublished - 2009
Event2009 Joint Urban Remote Sensing Event - Shanghai, China
Duration: 20 May 200922 May 2009

Publication series

Name2009 Joint Urban Remote Sensing Event

Conference

Conference2009 Joint Urban Remote Sensing Event
Country/TerritoryChina
CityShanghai
Period20/05/0922/05/09

Fingerprint

Dive into the research topics of 'Coherent stacking with TerraSAR-X imagery in urban areas'. Together they form a unique fingerprint.

Cite this