Robust SVA method for every sampling rate condition

Linear apodization, or data weighting, is the traditional procedure to improve sidelobe levels in a finite sampled signal at the expense of resolution. New apodization methods, such as spatially variant apodization (SVA), apply nonlinear filtering to the signal in order to completely remove sidelobes without any loss of resolution. However, the results are strongly influenced by signal sampling rate. Some variations which improve results have been previously published, but sidelobe cancellation gets worse since sampling frequency is not settled at Nyquist (or a multiple). This paper presents a new and efficient technique based on SVA that drastically reduces sidelobe levels for every sampling rate condition. The algorithm is, essentially, a parameter optimization of a variant filter for each pixel of the image. A one-dimensional case and a two-dimensional generalization are presented, as well as some applications to target detection capability in a synthetic aperture radar (SAR) system.