I am unable to get a decent stitch of these images with 2.5.0 A1 regardless of style of detection I use.

If I use Similarity then stitch is poor and I am unable to resolve the problems using the CP editor, if I fix one link then optimising removes an existing link and so on.

If I use Full 3D the result is even worse and the zenith image gets omitted altogether.

Image set from here:
http://www.chem.ox.ac.uk/oxfordtour/tut … 015mm%20FE

Download the images - shot with Canon 5D and Canon 15mm FE:
http://www.chem.ox.ac.uk/oxfordtour/can … w/jpgs.zip (36MB ZIP)

I am attempting to stitch after removing the two nadir images.

APG 2.0.9 has no problems at all achieving an excellent stitch.

First screenshot - APG 2.5.0 A1 using Ransac model - similarity.
Second screenshot - APG 2.5.0 A1 using Ransac model - Full3D.
Third screenshot - APG 2.0.9.

Last edited by mediavets (2010-06-25 14:55:05)