In this paper, we utilize a wave equation framework to describe SAR processing. Within this framework, various conventional and new processing techniques can be reliably compared. In particular, we compare wave equation implementations of the conventional range-Doppler algorithm and the seismic/SAR algorithm recently proposed. The range-Doppler algorithm is a matched filter solution of the Kirchhoff integral in range-time and Doppler-wavenumber space, while the seismic/SAR algorithm is a frequency-wavenumber or plane wave solution to the wave equation using the measured data as a boundary condition. Thus, both processors can be viewed as boundary value solutions to the electromagnetic wave equation. Point spread function analysis, of simulated ERS-1 data and of actual SEASAT data, shows that both processors provide very accurate relative amplitude and phase results. Nevertheless, we expect that future SAR processors may incorporate features of both techniques.

Satellite SAR interferometry estimates terrain elevation by comparing two complex SAR images collected from separate passes over common ground. We have implemented prototype software to perform satellite interferometry. Two raw SEASAT data sets from Northern Canada were used to test the software. This data presents an example of SAR interferometry in a northern area and also demonstrates the sensitivity of interferogram phase to the conditions of the underlying data. Complications to our analysis were introduced by the plethora of frozen and unfrozen waterbodies in the area along with varying weather conditions during the three day interval between data takes. Finally, irregularities in the data produced aninconsistent interferogram that could not be meaningfully transformed to height estimates. However, the generation of fringes from the Arctic imagery demonstrates that satellite interferometry is possible in near polar locations.