How to extract 2D backgrounds from video clips?

I want to rip the backgrounds from a cel-based animation.

In many cases, there is zooming, panning, rotation, etc. I'd like some way for it to stitch the pieces together to form the largest and highest res possible composite image. I don't mind needing to guide it a little, but the more automated, the better.

What software and/or methods can I use to accomplish this?

Nowadays there's pretty good image stabilization software that can do stuff like this with recorded footage. I have no idea how well that would work with something that was animated though.

Something like that would work great for what I'm hoping to do. What software was used in that linked video?