Abstract:

In order to enhance the spatial resolution of a single remote sensing image, a super-resolution reconstruction method based on sparse representation is presented in this work. First, a pair of dictionaries for low- and high- resolution image patches are learned using the majorization minimization method. The method substitutes the original objective function with a surrogate function that is updated in each optimization step, and can guarantee to find local minima in each optimization step. Second, given a low-resolution remote sensing image, the high-resolution image is reconstructed based on the pair of dictionaries.Our experiments show that the state-of-the-art results have been achieved compared to conventional interpolation methods in terms of both PSNR,SSIM and visual perception.The results demonstrate that our algorithm can provide useful high-frequency details for a single low-resolution remote sensing image in super-resolution reconstruction, and therefore the proposed method is universal.

Key words: Remote Sensing, Super-resolution Reconstruction, Sparse Representation, Dictionary Learning, Majorization Minimization Method