Super-Resolution of SAR Images With Speckle Noise Based on Combination of Cubature Kalman Filter and Low-Rank Approximation

In this article, two novel methods for synthetic aperture radar (SAR) image super-resolution (SR) are proposed. The main challenge for SAR image SR reconstruction is speckle noise. For this reason, a novel algorithm termed the importance sampling cubature Kalman filter (ISCKF) is proposed to reconstruct a high-resolution (HR) image from a series of low-resolution (LR) images. However, as the reconstructed image usually embraces residual noise visually, we establish a nonlinear low-rank optimization model in order to further reduce the speckle noise drastically. Correspondingly, an alternating direction method of multipliers based on the low-rank model (ADMM-LR) algorithm is proposed to solve it, which yields the other novel method termed ISCKF + ADMM-LR for SAR image SR. In addition, we establish the computational complexity of the proposed algorithms. The experimental results of both the simulated images and real SAR images demonstrate that the performance of the proposed methods is superior to some state-of-the-art methods in both SR and despeckle.