Two-dimensional extrapolation and spectral estimation from arbitrary sampling configurations for SAR/ISAR imaging

The adaptive weighted norm extrapolation (AWNE) procedure of Cabrera and Parks (1991) is generalized to non-uniform one-dimensional (1-D) and two-dimensional (2-D) sampling configurations. This extrapolation procedure which is based on iterative modification of the frequency weighting function in minimum weighted norm extrapolation produces high frequency-domain resolution. The approach presented is derived for totally arbitrary sampling configurations to include those that appear in inverse synthetic aperture radar (ISAR) imaging of a rotating object. The resulting procedure produces estimates of signal samples on rectangular coordinates in the neighborhood of the given samples (interpolation) as well as beyond (extrapolation). When applied to SAR/ISAR imaging, the procedure produces a complex reflectivity image in the slant-range cross-range domain from the samples of the object's Fourier transform, available at non-cartesian coordinate locations. Examples illustrate the performance of this interpolation/extrapolation procedure using samples of l-D and 2-D sums of complex sinusoids to simulate range profiles and images of a few point scatterers.<
>