 A new pan-sharpening method is proposed using convolutional neural networks that adapts a simple three-layer architecture for super-resolution and includes several maps of nonlinear radiometric indices to improve performance without increasing complexity. Experiments on three representative datasets show the method to provide very promising results, largely competitive with the current state of the art in terms of both full-reference and no-reference metrics, and also at a visual inspection. This article was authored by Giuseppe Massi, Davide Cotzilino, Luisa Verdeliva, and others.