Multitask Learning and Sparse Representation Based Super-Resolution Reconstruction of Synthetic Aperture Radar Images

In earth observing remote sensing fields, to recognize objects whose size approaches the limiting spatial resolution scale especially in Synthetic Aperture Radar (SAR) images, spatial resolution enhancement is usually required. In this paper, we proposed a multi-task learning and K-SVD based Superresolution image restoration method where K-SVD algorithm is employed to learn a redundant dictionary from some example image patches. In order to learn more accurate dictionary and reduce the complexity of dictionary learning, multitask learning concept is adopted to learn multiple dictionaries from the samples classified by K-means clustering. Some experiments are taken to investigate the performance of our proposed method, and the visual result and numerical guidelines both prove its superiority to some start-of-art SRIR methods.