I need to georeference 53 air photos from 1981 (1:25,000 scale) and was wondering what would be the most efficient, defensible, and accurate option. For context, the air photos are following a highway taken along a valley bottom and therefore the relief would not be significant as would be the case if mountains were involved. Essentially, I want to quantify the change in surface area around this highway from 1981 photos to current day; how do I process the former images?
Typically, one would georeference each air photo using current satellite imagery. However, in my case, I feel this would be difficult because the area of interest has changed through time, and therefore I fear I will not have enough reference points from current satellite imagery to georeference each air photo from 1981. Therefore, I thought I could first stitch the 53 air photos together using control points from the adjacent air photo taken the same day; essentially creating a mosaic from 1981. Then I could georeference the entire mosaic using current satellite imagery; I think on a larger scale I could find enough reference points to accurately geofrerence the mosaic. Would this process cause too much local distortion that I would no longer be able to accurately quantify surface area? Lastly, I'm curious if orthorectifying the image would significantly improve the surface estimation? I'm not looking for complete precision, but would like it to be accurate enough for a publishable manuscript.