Robustness of SAR Refraction Autofocus to Power-Law Errors

A recent analysis has yielded a non-parametric methodology which automatically estimates and compensates for atmospheric refraction defocus in synthetic aperture radar (SAR) imagery. Such deleterious defocus effects result from the delay and bending of the individual SAR radar pulses as they refraction through the atmosphere. A primary benefit of the recent data-driven refraction autofocus is that it is robust to non-parametric delay and bending errors which can vary from one pulse to the next, as arises due to changing atmospheric conditions along the synthetic aperture of the radar's trajectory. The current investigation examines the robustness of these techniques for realistic scenarios in which the refraction-induced delay and bending vary in time according to a power-law spectrum. Specifically, this refraction autofocus methodology yields sharp scene focus when applied to measured Ku-band SAR imagery in which power-law errors have been injected into the delay and bending of the radar pulses.