ISAR Imaging and Motion Compensation with Sparse Bayesian Learning

In high-resolution inverse synthetic aperture radar (ISAR) imaging, the translational motion of a target introduces phase error in the echo data and its correction is an important issue. Besides, the rotational motion of strong maneuvering target also introduces time-variant Doppler modulation in short coherent processing interval (CPI) and is necessary to be compensated for well-focused imaging. In this paper, we focus on high-resolution ISAR imaging joint with translational and rotational motion compensation from data of compressed sampling. A chirp-Fourier dictionary is used to represent the maneuverability of rotational motion. Meanwhile, the translational phase error is treated as model error in ISAR imaging. Then, ISAR image formation is constructed by using a hierarchical statistical model to encode a sparsity prior and solved from sparse Bayesian learning, including sparse image reconstruction, parameter estimation and phase error correction. As a result, non-ambiguous imaging with accurate motion compensation can be achieved even by using the data of compressed sampling. Finally, experiments are performed to confirm the effectiveness of the proposed method.