To solve this issue, we propose a data augmentation framework for pansharpening neural networks (PNNs). Specifically, we introduce first a random spatial degradation based on anisotropic ...

Those missing entries which make valuable information invisible may distort the further urban analysis. To improve the quality of missing data imputation, we propose an improved spatial multi-kernel ...