CloudCompare forum

I've been experimenting with a workaround in CC to align scan data from a handheld scanner to a photogrammetry-derived mesh. It's straightforward setting control points, assigning the reference, and running the refined alignment tool afterward. I encountered an issue numerous times after a decent alignment with control points where the mesh being aligned and scaled would return seriously bad scale, like less than half the size of the reference. Anybody encountered that?

This test involved a very low brow bit of hardware, iPhone 13 photography (12 MP imagery) and unordered scan data from BLK2GO, which looks way soft to my eye. (I'm now switching to BLK360, but that's beside the point.) Might the poor quality data be the culprit here?

Hum it shouldn't?

But what generally leads to this kind of issue is when too few pairs are taken, and mostly aligned (along a line, or even in a plane).

Got it. So, if I'm following, the control points should be along all three axis and spread out as far as possible to avoid rotating around an axis or smaller region, yes?

Yes, that's always the best way to do it (if possible).