Phase Noise Countermeasures for Synthetic Interferogram Generation

During synthetic interferogram generation (whereby a DEM and platform positions are used to calculate expected fringes), an interpolation step is necessary during the transformation from map geometry into the slant-range geometry of the radar image. Particularly in hilly terrain, assumptions made during the interpolation step can have an impact on the accuracy of the simulation. We investigate differences between simulations made using differing assumptions and interpolation methods, and provide algorithm recommendations. We show results both from one-pass methods (where no knowledge of each radar geometry grid location’s neighbouring slopes is assumed) and a two-pass method (improved accuracy at the price of increased computational cost). In the latter method, after the initial pass through the DEM, a second-pass rangesorted traversal is made (allowing knowledge of each point’s immediate neighbours, even in hilly terrain) before the final interpolation for each point. We highlight the effects of differing algorithms for various terrain types, and discuss how assumptions about the extent of topography within a scene can be used to accelerate generation of the synthetic image products. We demonstrate the differences using simulated ASAR imagery produced using ERS data.