Signal Processing : Services
CSL developed a recognized expertise in the frame of SAR interferometry (InSAR), differential interferometry (DInSAR), coherence tracking, SAR polarimetry and polarimetric interferometry (PolSAR & PolInSAR), split band interferometry (SBInSAR) and spectral coherence.
SAR interferometry (InSAR) consist in using the phase information in two consecutive or simultaneous SAR acquisitions of a given area to generate an interferogram and a coherence image of the observed scene. Main InSAR application consists in retrieving a Digital Elevation Model (DEM) from the interferogram. Coherence is a quality indicator of the scene stability between acquisitions. It can be used as an indicator of backscattering processes changes. Figure 1 shows a DEM of the Nyiamuragira volcano overlaid with a false color image in which both the original SAR image and the coherence image were used.
Figure 1: InSAR DEM of the Nyiamuragira volcano
Coherence tracking consist in optimizing coherence at a local level to refine co-registration of images used in the interferometric process. Such tracking allows to detect azimuth fast moving features between acquisitions. Animation hereafter shows coherence tracking results as observed for the Shirase fast flowing ice stream in Antarctica.
If the topography of the scene is made available, the topographic phase may be retrieved from the interferogram to show only the phase component linked to local displacements. This process is know as differential interferometry (DInSAR). Figure 2 shows DInSAR result derived for the Landers earthquake (1992).
Figure 2: DInSAR fringe pattern of the Landers earthquake
Two SAR images, one before and one after the earthquake, where used to generate an interferogram containing both the topography and the displacement. A second interferometric pair of images was used to generate only the topography onto which the differential fringe pattern was overlaid.
Polarimetric SAR (PolInSAR) make use of polarimatric channels if available in the SAR acquisition to perform target decomposition and segment the signal along three main backscattering processes. Full polarimetric SAR data allows also performing polarization synthesis. The animation hereafter shows an example of polarization synthesis showing what would be observed if sending a left circular polarization while analyzing along variable horizontal polarizations.
Polarimatric SAR interferometry
Polarimatric SAR interferometry (PolInSAR) uses target decomposition principle to perform coherence optimization at local level. This optimization results in three interferograms, each corresponding to a dominant backscattering process. Figure 3 shows an example of optimized interferograms in line with their associated optimized coherence.
Figure 3: Polarimetric SAR interferometry
New sensors are using wide bandwidth signals to improve resolution. This wide band may be used as a new degree of freedom. Splitting this wide band in sub bands allows performing a spectral analysis of a single image, leading to sub-images of lower resolution but centered on slightly different carrier frequencies. Performing a coherence analysis between sub-bands allows computing spectral coherence. Manmade structures are shown to be spectrally stable while random distribution of scatterers are loosing spectral coherence. This characteristic allows for example detecting vessels on open sea surface. Figure 4 shows an example of spectral coherence measured on the Venice Laguna.
Figure 4: Spectral coherence measured on the Venice Laguna
Split-Band SAR interferometry
This sub band splitting process may be applied to an interferometric pair of SAR images leading to a spectral interferometric decomposition. This technique know as Split Band Interferometric SAR (SBInSAR) allows measuring the "absolute" interferometric phase on a point by point basis, alleviating the well known phase unwrapping problem. Figure 5 shows an example of a DEM of the Uluru monolith (Australia) fully obtained using SBInSAR.
CSL InSAR Suite
Contact(s) : Dominique De Rauw firstname.lastname@example.org