Need expert! Strange data issue! Has Agisoft changed its stitching algorithm?

[sanrory]

Hi all,

We are having a really unusual issue with agisoft that we cannot understand.

Attached are two different images. the aprfeb.jpg image is a height map comparison between two sequential data sets, april and february, and the febjan.jpg is the same comparison across different months.

there are ~20 gcps around the site and they have not changed at all. Jan Feb and Apr are all complete datasets, 70% side and overlap.

The issue is with the april and february comparisons, the z heights all match up at the GCPs, but on either side we are seeing crazy changes which really dont match up at all. There is a road which lines the site in the south and that is skewed as well.

Literally all of the processing settings are the same, the GCPs are the same, does anyone know what this could be?

Thanks

[MartinBim]

So your scale bar goes from -5 to 5 and from -3 to 3 in the second image. Did you checked this?

[sanrory]

Yes the problem is there regardless of the scale... on the left hand side of both photos there is a big cement slab with brick pallets on it, the slab height hasnt changed at all between Jan and Feb but between Feb and Apr it apparantly has changed significantly.