ISAR Imaging using Adaptive ARMA Spectrum Estimation in Subbands

The objective of this paper is to present an Inverse Synthetic Aperture Radar (ISAR) image formation technique suitable for radar target recognition tasks. The one-dimensional (1-D) input data is first interpolated, converted to two-dimensional (2-D) format, windowed, and modeled as 2-D autoregressive moving average (ARMA). After zero-padding, it is splitted into subimages using the 2-D filters generated from a (1-D) prototype filter, and decimated along each dimension before once again converted to 1-D multichannel signals in order to adapt prediction filter coefficients prior to ARMA spectrum estimation. The sequential processing multichannel lattice stages (SPMLSs) are used as adaptive prediction filter structure, and conversion of lattice parameters to process parameters is accomplished using a Levinson-Durbin type conversion algorithm specifically developed for SPMLSs. The performance is Spectogram,visually presented, which and is the compared squared against magnitude the performance of the short-time of the Fourier transform (STFT).