How different can two reasonable pansharpening algorithms be? Pretty different in pathological cases like planes in flight. Left is from Maxar, right is from Potato (the thing I’ve been tinkering with). Same underlying data, from the Maxar Open Data Program, CC BY-NC.
@kissane ghost_detector.py
@vruba are there any good walkthroughs of going through this process with open-sourced tools? I'm fascinated by the process but I don't know where to start without paid tools.
@boxly If you find one, tell me? I’m going to open-source my code soon, I hope. Unfortunately pansharpening is really under-theorized. People mostly do it without talking about it in detail. There are research papers on arxiv, some with associated code, but to be blunt most of them are not very good. (Lots of undergrad-written Matlab.)
@vruba thanks for the reply! I will look forward to anything you publish and will have to start digging into some papers I don't really understand 😅
All we need is a long vehicle
@vruba oh you found a way to detect the lag-ghosts, good.