Superresolution of remotely sensed images with anisotropic features

We consider the problem of superresolution for remotely sensed images. Our ambition is to develop an algorithm that efficiently increases the resolution of an image without introducing artifacts or blurring, and without using any additional information, such as images of the same scene in different modalities or sub-pixel shifts of the same image at lower resolutions. The approach developed in this article is based on analysis of the directional features present in the image that is to be superesolved. The harmonic analytic technique of shearlets is employed in order to efficiently capture the directional information present in the image, which is then used to provide smooth, accurate images at higher resolutions. Our algorithm is compared to the standard superresolution method of bicubic interpolation. We test our algorithm on a remotely sensed image of Gulfport, Mississippi. Our results indicate the superior performance of shearlet-based superresolution algorithms, compared to bicubic interpolation.