This is a collection of notes on how I’ve been approaching CNNs for pansharpening. It’s an edited version of an e-mail to a friend, who had asked about this tweet, and so it’s informal and somewhat silly; it is not as polished as, say, a blog post would be. Basically, what I’ve written up here is the advice I would give to an enthusiastic image processing hobbyist before they started working on pansharpening.
If you want a more serious brief review of the problem, try the first two pages of Learning deep multiresolution representations for pansharpening – most of the work I would recommend is mentioned there.
If you spot an error in the physics or math here, please mention it, but if you see a simplification or minor abuse of terminology, it’s on purpose; for textbook rigor, please refer to a textbook. The focus here is