On 28/01/12 12:33, kfj wrote:
On 27 Jan., 18:28, Doug<[email protected]> wrote:
I haven't bothered to remove sky points because it hasn't
seemed to make any difference with these images
If I was you, I'd get rid of all points in the sky. Misalignment in
the clouds is hard to spot.
I thought to try assigning separate lenses to each stack and
unlinking the lens parameters, then optimising for all lens
parameters including fov. The result was one step forward,
two steps back - the distance is now fine over the whole
panorama, but the foreground register is worse, even
extending a bit beyond the centre.
What would you suggest next ?
Assuming all options in hugin have been used to the max, there is just
one solution left over: redo the take with a tripod - assuming the
errors you have are from parallax. Exposure stacks are very sensitive
to parallactic errors. Why so? In a standard panorama, the images only
need to match along the seams, and the intelligent seam placement can
put the seam in a place where the match is good. When fusing stacks,
every single pixel from every single image is taken into account
(within the constraints of the fusion parameters) - so if there are
mismatches, they will show. If a tripod is not an option, do at least
use a stick (like, a walking stick') to rest the camera on, and if all
else fails, rest it against a tree, a rock or your body. Do the
bracket in serial shot mode, so as to give you as little chance as
possible to move the camera between shots.
Ther's one thing you may want to try in case you can't redo the take.
Often it's the case that parts of the image don't need to participate
in the final outcome. If this is true for the parts of your image set
where you experience problems, you can do this:
Switch off all images but one from each stack, so that the only images
left are the ones which have the best exposure for the area where your
fusion problems are. Alternatively you can just switch off those
images which obviously can't align well with the others. Now stitch a
panorama from the reduced set (make sure you don't optimize or change
anything before creating the subset panorama). Now you can open the
fused panorama you already have and the new subset panorama as layers
in an image editor and use a layer mask to override the erroneous
areas in the fused panorama by parts of the subset panorama.
Kay
Thanks again Kay for your most informative advice.
Meanwhile I've tried a last resort, chickening out of doing
everything at once, and settled for fusing each stack as an
individual hugin session (having removed sky points) -
obviously not very practical if I'd had a lot of stacks
instead of only six! - and then running a separate session
to blend the fused images together. That finally produced a
decent result.
Five of the six stacks fused individually as simple
panoramas without problems, but the sixth - on the
right-hand edge of the picture, much to my surprise, as I
would have expected the problem to lie with the middle
stack(s) - took quite a bit of effort to bring it into good
register. I had to unlink the images, set each image a new
lens, and optimise for y,p,r,X,Y,Z and v (having tried
optimising fewer parameters first unsatisfactorily).
The three bracketing images would have been taken within a
fraction of a second of one another, seeing that the camera
was set to AEB and continuous shooting, so I can't imagine
what would have been able to go wrong in such a brief interval!
Doug
--
You received this message because you are subscribed to the Google Groups "Hugin and
other free panoramic software" group.
A list of frequently asked questions is available at:
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at http://groups.google.com/group/hugin-ptx
