Remote sensing image super-resolution reconstruction algorithm based on improved anchored neighborhood regression

In recent years, the application scenarios of satellite remote sensing images have become increasingly widespread. However, due to limited acquisition equipment and cost constraints, the images obtained by satellite sensors usually don’t have sufficient resolution to meet the analysis and discrimination research of subsequent satellite images. Based on the ideas of neighbor embedding and sparse representation, the super-resolution reconstruction method of remote sensing images using Adjusted anchored neighborhood regression (A+) is studied in this paper. This algorithm is based on the sparse representation theory to train a dictionary. With the dictionary atoms as the center, K feature patches that are most similar to them are selected to form a neighborhood, and then the projection matrix corresponding to this neighborhood is calculated. When reconstructing, the atom that is most similar to the low resolution feature patch is selected, and the projection matrix corresponding to this atom is used to obtain the reconstructed high-resolution image. This article compares A+ with other methods based on sparse representation through simulation experiments, verifying the effectiveness of the algorithm.