A New Public Alsat-2B Dataset for Single-Image Super-Resolution

Recently, deep learning methods dominate the proposed solutions for image super-resolution due to their powerful properties. However, for remote sensing benchmarks, it is very expensive to obtain high spatial resolution images. Most of the super-resolution methods use down-sampling techniques to simulate low and high spatial resolution pairs and construct the training samples. As an alternative, the paper introduces a novel public remote sensing dataset (Alsat-2B) of low and high spatial resolution images (10m and 2.5m respectively) where the high-resolution images are obtained through pansharpening. Besides, the performance of some state-of-the-art methods is assessed based on common criteria. The obtained results reveal that the proposed scheme is promising and highlight the challenges in the dataset which show the need for advanced methods to grasp the relationship between the low and high-resolution patches.