2-D sparse autoregressive modeling for high resolution radar imaging

ISAR imaging based on autoregressive (AR) model has not only spurious scattering centers but also high side lobes. Sparse AR models can be utilized for suppressing these. However, computational complexity of the BPDN with penalty sparsity approach which is employed to compute sparse AR model coefficients is high. In this work, the sparse AR model coefficients are computed by using BPDN and LASSO approaches which have less computational complexity. Spurious scattering centers and side lobes are successfully suppressed in the resulting radar images.