| Popis: |
Downward looking linear array three-dimensional synthetic aperture radar (DLLA 3-D SAR) has been widely used in digital surface model generation, planning and environmental monitoring, object detection, identification and tracking, etc.. While for practical system, the equivalent antenna phase centers (APC) always suffer from loss or deviation, leading to the sparse and non-uniform distribution of APCs. Since 3-D imaging scene usually behaves spatial sparsity, compressive sensing (CS) based sparse recovery methods are available for DLLA 3-D SAR imaging. However, traditional CS methods, which assume the scatterers located on the discretized grids, will suffer from off-grid effect and basis mismatch problem when true scatterers deviate from the grids. In this paper, we propose an imaging method for DLLA 3-D SAR combining back-projection and an improved orthogonal matching pursuit (OMP) approach. 2-D focused images are firstly obtained by back-projection algorithm with Fraunhofer approximation along cross-track dimension to reduce the phase error. Then the improved OMP approach is applied in each along-track and wave propagation resolution cell to reconstruct the cross-track dimensional scatterers distribution. By solving a joint-sparse recovery problem, the proposed method can give off-grid effect-free reconstructed results. Experiments from simulated data and real data demonstrate the validity of the proposed method. |
