Sparse-based reconstruction of missing information in remote sensing images from spectral/temporal complementary information

Because of sensor failure and poor observation conditions, remote sensing (RS) images are easily subjected to information loss, which hinders our effective analysis of the earth. As a result, it is of great importance to reconstruct the missing information (MI) of RS images. Recent studies have demonstrated that sparse representation based methods are suitable to fill large-area MI. Therefore, in this paper, we investigate the MI reconstruction of RS images in the framework of sparse representation. Overall, in terms of recovering the MI, this paper makes three major contributions: (1) we propose an analysis model for reconstructing the MI in RS images; (2) we propose to utilize both the spectral and temporal information; and (3) on this basis, we make a detailed comparison of the two kinds of sparse representation models (synthesis model and analysis model). In addition, experiments were conducted to compare the sparse representation methods with the other state-of-the-art methods.