2 - SAR Imaging using Coherent Modes of Diversity: SAR Polarimetry, Interferometry and Tomography

Abstract

Spectral and spatial diversity measurements, Synthetic Aperture Radar (SAR) imaging uses modes to focus 2D images of electromagnetic reflectivity of environments. Even though this information is very useful for detecting objects or to evaluate some of their geophysical properties, it has some limitations for further applications or characterizations. In fact, very diverse objects, such as a vehicle or a forest parcel, may provide responses with a similar energetic level. As a result their differentiation must necessarily involve additional information, among which polarization diversity is often considered. Classic SAR imaging uses a system of antennas polarized in the same way, and so measures 2D scalar information. The use of antennas with different polarizations allows the measurement, for each image pixel, of a multi-variate polarimetric quantity which provides information on the intrinsic geophysical properties of the imaged objects. Another significant limitation of SAR imaging comes from the natural ambiguity linked to the 2D mapping of 3D environments. This dimension reduction causes a direct loss of height related information because of height and, in particular, mixes the contributions from scatterers located at different altitudes. The solution usually considered uses additional spatial diversity to determine, for each 2D SAR image pixel, either the elevation position of the main scatterer, or the vertical distribution of the imaged reflectivity, for tomographic applications. It should be noted that SAR interferometry (InSAR) requires the coherent combination of N=2 SAR measurements, compared with N≥3 for tomography and 2≤N≤4 for polarimetric measurements. These different measurement modes, which can be combined with each other, are illustrated. The use of wave polarization or the measurement of an object’s position through interferometry are widely known techniques used since the post-war period in the field of optics. During the 1950s and 1960s, a group of researchers developed the theoretical tools necessary to handle coherent multi-variate data, using wave polarization to differentiate radar obstacles or to identify them from their unique polarimetric behaviors. In the 1970s, deeply modified the approach of polarimetric data processing by combining a rigorous mathematical analysis with a phenomenological interpretation of electromagnetic scattering mechanisms, paving the way for sophisticated techniques such as the ones presented in this chapter. SAR interferometry, which can determine the elevation position of the different SAR image pixels, appeared in the 1970s–1980s with airborne applications, and then satellite applications with SEASAT-A measurements. The introduction of a nearly uninterrupted series of satellite missions adapted to the InSAR mode since the ERS-1 launch in 1991 has resulted in a boom in SAR interferometry techniques, which have now reached a nearly industrial level of maturity and which provide, through differential SAR interferometry, a technique with unique performances. SAR tomography was born at the beginning of the 2000s and has made it possible to image a scene in 3D by combining several inSAR images. To date, its application is limited to airborne data, but adjustments are currently being studied to extend this approach to satellite measurements.