How to implement a comparison between 2 sets with a large number of flight / satellite photos?

Question

Automation

Good day!
Subject: There are "n" number of photos from a certain drone route (monitoring, geodesy).
Task: To automate the process of comparing a set of photos of a new flight with the previous one, the photos will not be identical in frames (gusts of wind, etc.), that is, as I understand it, you first need to associate the photo of the new set with certain coordinates generated from the previous set and then conduct a comparative analysis . And so every time, new with the previous one. How to implement? Neural network?
Thank you!

Answer 1

beyond compare ?

Answer 2

Classical photogrammetric chain - get orthophoto, then compare pixel by pixel (within adjustment accuracy). Otherwise it will be trash.

Answer 3

And it was impossible to tie the coordinates to the photo number in advance?