Difference of DEMs Error

[Stefano]

Hi All,
I made 2 drone surveys, the second after one year on a glacier with a on-board RTK system and (unfortunately)
with different cameras. I used some markers as reference points for the reprojection error and not for the
bundle adjustment since I have georeferenced cameras. RMS reprojection error is very good, so as camera errors (m)
and markers (px). The orthophotos I obtained are good but DEMs are very strange because the DEM of difference
indicates stable areas (rock outcrops) with 1 to 4 meters difference. Subsequently, I tried to use the estimated
markers coordinates from the first model as GCP for the second model. So I rebuilt the second model with
GCPs and no camera references. Once again I obtained good errors but at the moment of the DEM difference
the results were the same as before.
Does anyone know what I did wrong? I would like to upload the project but it is very big.
Thank you all in advance.
Stefano