Dynamic ISAR Imaging and Autofocusing of Maneuvering Targets Based on Sequential GP-SOONE Method and Eigenvalue Decomposition

A long coherent processing interval (CPI) is needed for achieving a high-resolution inverse synthetic aperture radar (ISAR) image. However, for a maneuvering target, the time-varying Doppler shifts cause a blurring effect on the ISAR image. Sparse representation-based algorithms can obtain a high-resolution image in a short CPI, while the Doppler shifts remain constant. Recently, a sequential order one negative exponential (SOONE) function has been introduced to measure the sparsity, and a gradient projection (GP) method has been used to solve the SOONE function and recover the sparse signal. In this paper, a 2D sequential GP-SOONE method for sparse recovery and dynamic ISAR imaging is proposed, which has a lower computational complexity than that of the 2D-GP-SOONE algorithm. Moreover, the performance of the proposed approach is the same as the 2D-GP-SOONE and better than the 2D smoothed L0 algorithm. Another problem of dynamic ISAR imaging is sequentially autofocusing the image. Hence, a fast parametric method based on eigenvalue decomposition and minimum entropy for dynamic ISAR autofocusing is proposed which has a faster convergence than the conventional methods. The proposed method has also comparable performance with the conventional ones. Several simulations and real data are used to show the superiority of the proposed methods.