I don't thing this is an issue, as long as there isn't a pattern in the background that does not match to the position of the object, for instance, if you turn the object itself and Meshlab align things in the the background.
I am currently experimenting with featureless background (dark),
to be able to add the bottom of the object, so that the 3D model
is visible from all angles, and so far, it usually work. The worse
case for me is when Meshlab does not have enough features to link
photos from the bottom and the rest.
I have not tried transparent background though, but I did but my
objects on a piece of glass to avoid shadows on the background.
Is it possible to scan an object if all the pictures have a transparent or solid color background?
I have similar pictures, and I can't get pass the StructureFromMotion node. I assume it's because my background has no detail or interesting markers.
--
You received this message because you are subscribed to the Google Groups "AliceVision" group.
To unsubscribe from this group and stop receiving emails from it, send an email to alicevision...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/alicevision/9a7319af-680b-475d-b4ee-bd155a47b912n%40googlegroups.com.

Is it possible to scan an object if all the pictures have a transparent or solid color background?
I have similar pictures, and I can't get pass the StructureFromMotion node. I assume it's because my background has no detail or interesting markers.