scattering matrix available for analysis. Interferometry relies on the relative phase stability of the scene for determination of digital elevation models.
Interferometric SAR is a technology that uses the phase difference between two coherent synthetic aperture radar images of a scene, obtained by two receivers separated by a cross-track distance called the baseline, to measure the height of the imaged surface. The interferometric technique was first applied to SAR by Graham (1974), who used two slightly displaced antennas mounted on an aircraft to form an interferometric beam (Figure 2.1a). When two antennas are used to receive simultaneously the complex signal scattered by the scene, subsequent to illumination by one of them, and then are processed together to produce height information, such a scheme is called cross-track SAR inteferometry (see Box 2.1).
Seasat in 1978 and SIR-B in 1984 generated many SAR images of Earth 's surface. Goldstein et al. (1988) extended the two-antenna interferometric SAR technique to a new capability by demonstrating how multiple-pass Seasat SAR images can be combined to realize the interferogram from which height information is derived (Figure 2.1b). Improvements in oscillator phase stability allowed Zebker and Goldstein (1986) to generate three-dimensional maps of large areas at resolutions that are practical for many topographic applications. A similar demonstration was conducted using SIR-B images (Gabriel and Goldstein, 1988). In the 1990s the technique has been applied to numerous multiple-pass image combinations recorded by ERS-1 and 2, JERS-1, and Radarsat.
Another recent development uses a full-polarization matrix, including phase, to obtain information not easily acquired in other ways. Much research since the 1980s shows that for some applications, phase differences between the polarizations convey information on the structure of the surface or of vegetation. Although radars of the 1960s had multipolarization capability, they could not measure the phases between polarizations and thus could not provide information conveyed by the phase. Rather, they provided only the amplitude differences, with those between like- and cross-polarizations being most useful.
A small SAR system design must consider not only parametric and cost trade-offs, but also issues involving applications, data access, and data dissemination. A well-conceived end-to-end system should include applications undertaken for the public good as well as those driven primarily by science and commerce. Many of the potential applications cited by Evans et al. (1995) would be beneficial to multiple user sectors but are viewed differently by each of them. Intuitively, some applications are oriented more toward a single user sector. However, most