A Novel Airborne Tomographic SAR Reconstruction Method Based on Sparse Bayesian Inference

Advancements in synthetic aperture radar (SAR) technology have enabled the development of 3-D imaging as a significant application. Tomographic SAR (TomoSAR) enhances the SAR sensor’s imaging performance by synthesizing a 2-D virtual aperture in the altitude direction, thereby enabling 3-D imaging. Imaging along this direction in TomoSAR is fundamentally a spectral estimation problem. Creating synthetic apertures in the altitude direction, however, presents greater challenges compared to the azimuth direction. Consequently, the length of apertures in the altitude direction is considerably shorter than in the azimuth direction, leading to a lower resolution in the altitude direction if reconstruction is performed using the Fourier transform (FT) method. Given the constraints on aperture resources, there is an urgent need for effective high-resolution imaging algorithms for the altitude direction to achieve high-resolution 3-D imaging. This article investigates high-resolution, 3-D TomoSAR imaging within the framework of compressed sensing (CS). A modified imaging algorithm based on off-grid sparse Bayesian inference (OGSBI) is proposed. The algorithm employs OGSBI for spectral estimation, integrating sparse extraction and generalized likelihood ratio test (GLRT) selector to enhance the performance of the imaging process. The proposed algorithm achieves high-resolution 3-D imaging, demonstrated through simulations and processing of measured data, even under conditions of limited synthetic aperture length in the altitude direction and nonuniform baseline spacing. The correctness and effectiveness of the method are validated through these evaluations.