Sparse Decomposition-Based 3-D Super-Resolution Imaging in Millimeter Wave SAR

Three-dimensional imaging capability can be obtained in millimeter wave array synthetic aperture radar (SAR). The resolutions are, however, limited by the bandwidth of the transmitted signal and the length of apertures, which, in turn, affects the performance of 3-D modeling. In order to address the above problem, this article proposes a sparse decomposition-based 3-D super-resolution imaging method. The signal model of the received back data is formulated and a modified range Doppler algorithm (RDA) is proposed to realize 3-D imaging. Based on the modified RDA, a group of azimuth filters are calculated to construct a projection dictionary. The received SAR data is viewed as the sparse representation of the projection dictionary, where sparse and image difference feature constraints are employed on the solutions. After iterative calculation, the mixed back data reflected from targets with different locations will be separated and decomposed gradually, resulting in super-resolution imaging performance. Finally, numerical simulations and real millimeter wave SAR data processing are performed to verify the effectiveness of the algorithm.