I'm learning as I go here ... but I think there are two separate things happening. One is the vignetting (which is a consistent variation across each individual image.) And then the other is the overall exposure (or white balance or gamma?) There is probably some super smart way to do a global optimization to match all of this up in one step? My little experiment only focused on the vignetting bit. I've also seen people do histogram matching (i.e. force/interpolate the histogram of one image to match the histogram of another image.) In a set of 2000 images with widely varying content, which image histogram should be the master? Also, histogram matching wouldn't correct for vignetting.
What I have been doing after the vignetting step is an adaptive histogram equalization (on an 8x8 grid). I like this because it makes the shading details of the image really pop out and it levels out the exposure differences (but still doesn't account for white balance differences.)
The dataset from the original post was captured about an hour before sunset on a thick/cloudy/dreary winter afternoon, so the lighting was about worst case. For this data set the auto white balance (or gamma?) of the camera really started showing through. For typical well lit sunny days the differences are much less noticeable. But maybe some sort of histogram matching could account for all of that?
For what it's worth, I've flown with both full manual mode and shutter priority mode. I flew an area of steep terrain in January with full manual setup and once side of the canyon was perfectly exposed and the shaded side was very under exposed. So then I started experimenting with shutter priority so I could at least control exposure time, but let the camera pick iso and aperture. I'm usually out imaging wooded areas with terrain so it has been really tricky to get the manual settings exactly right. (I do fly with fixed manual focus of course, but even managed to screw that up on a flight or two.)
I was wasn't so much asking a question here, but more sharing my experiments. But always enjoy feedback and new perspectives. I just see one tiny slice of the world from one tiny perspective. I haven't ventured over to the open-drone-map forum much because I figure they prefer to talk about open-drone-map there and not other people's projects.