A gridless sparse ISAR imaging method using atomic norm minimization based on multiple measurement vectors

The atomic norm minimization (ANM) based gridless recovery approaches can completely obviate the grid mismatch problem of discrete compressed sensing methods by working directly on a continuous dictionary, which have attracted considerable interest in sparse inverse synthetic aperture radar (ISAR) imaging. In order to exploit the joint sparsity in the multiple measurement vectors (MMV), a MMV-ANM approach for sparse ISAR imaging with stepped frequency signal is proposed in this paper. By reformulating the sparse range frequency echoes into a MMV-ANM model, the expected full echo without off-grid can be recovered at a single step by solving a semidefinite programme (SDP). Finally, the further improved ISAR imaging results can be achieved via the standard fast Fourier transform methods. The real data experiments demonstrate performance of the proposed method compared to existing gridless approaches.