Hi everyone,
I'm testing a workflow to scan a set with photoscan, and my friend and I are used to undistort photos before to use them in tracking softwares. To do so, we use Enwaï, a standalone and maya plugin that allows us to use a specific grid of distorsion for our specific cameras, with a lot of options (non linear undistort process, etc..). It's very powerfull and the result is very accurate, so we thought that it could be better to undistort photos before using them in photoscan.
But it seems that it doesn't work well if we give to photoscan undistorted photos to align.
I just want to be sure of the full process you recommend.
Do we just have to give photo (with distorsion) to photoscan, and then export undistorted photos after the align photo process, in order to camera map the model?
If this is the good way to use photoscan, what about your undistort process?
No offense, but I'm sceptical about the full automatic process, is it really as accurate as a process with a specific distorsion reference grid?
C.