Image Spectrum Decomposition of Ice-Sounding Data in Stratified Medium for Back Projection Algorithm

Image formation of synthetic aperture radar, which virtually observes the scenes and identifies targets along the predetermined trajectory, can be applied to ice-sounding data in the time domain, to map subglacial topographic structures with less sidelobe clutters. However, complex signal transmission path within stratified media limits the efficiency and accuracy of focusing. Here, we present the method to analyze the wavenumber support region by decomposing in the spectrum domain the combined vector formed from multi-paths in stratified media. This assists to determine the lowest Nyquist sampling rate, which essentially accelerates the reconstruction of aliased-free images and can be inherently inserted into the fast factorized back projection algorithm framework, based on divide-and-conquer scheme, recursively yielding low-to-high quality images. With this technique, along-track motion perturbations of flights and unknown tomography variations, can be accommodated at the additional but neglectable computational expenses. Results from real ice-sounding data in Antarctica are presented to validate the promising performance of the proposed method.