3-D Sparse SAR Imaging Based on Complex-Valued Nonconvex Regularization for Scattering Diagnosis

Linear array synthetic aperture radar (SAR) can achieve 3-D high-resolution imaging, which provides a novel scattering diagnosis technique. However, high sidelobes and background noise are the main challenges of scattering diagnosis based on SAR image. Sparse imaging can improve image qualities, such as suppressing sidelobes and noise. L-1 regularization is an efficient and typical model for sparse imaging. However, the L-1 regularization, as a convex optimization method, often introduces bias in amplitude estimation, which has a negative impact on scattering diagnosis. Therefore, in this letter, a 3-D imaging method based on complex-valued minimax concave penalty (CMCP) and improved alternating direction method of multipliers (IADMM) is presented to obtain high-quality and high-accuracy 3-D SAR images for scattering diagnosis. Compared with the existing sparse imaging method based on L-1 regularization, the proposed method not only improves the image quality, but also reduces the bias effect, which can be applied for scattering imaging. In addition, IADMM significantly reduces computational complexity. The experimental results indicate that the proposed method has compelling reconstruction accuracy of SAR scattering image.