BugBear wrote on Wed, 22 Oct 2014 at 10:08:30 +0100:
----------------
jepz11 wrote:
>
>
> On Tuesday, October 21, 2014 12:44:13 PM UTC+2, bugbear wrote:
>
> to keep one of the images (that happens to carry more important
> subject matter) untransformed, and get hugin to calculate and apply
> transforms only to the second image.
>
>
> Use a mask 'include region' for the whole first image?
I need the second image to be transformed such that it is a dead fit
for the unaltered first image; I have every intention of using
the source image unaltered for the first image, so it doesn't really
matter what hugin does to it. Only getting the right transform
for the second image (somehow) matters.
---------------
You tell us you don't want to transform the first image at all.
Therefore, use it as the anchor, with all its parameters zero and
non-optimisable.
Keep the lens parameters as imported, with, in the example I tried, just
an fov and use a single lens.
Optimise the second image for yaw, pitch and roll. That superimposes
the second image on the first as they sit in Hugin.
Now switch to the Stitcher tab and click Calculate optimum size (it may
be possible to do so earlier, but I haven't tried). Having done that I
reoptimise and normally get a bit of a shock at the greater errors
announced by the optimiser now the output size is greater.
I applied the above to a couple of group photos taken from the same spot
and then got Hugin to output its version of the anchor image. I found it
superimposed pixel-for-pixel on the original, though with a fringe where
the size did not quite match because I cropped within Hugin. In the
group photos I then applied masking to select the heads I preferred
(giving the same effect as using the fuse function in Microsoft Photo
Gallery, if I dare mention that). But combining the images in an
editor, as I think you intend, is probably the better approach.
Nonetheless it may be worth outputting both images from Hugin and
working with those because they will already be in register.
Trivial improvements in optimisation could be achieved by allowing other
parameters to vary for the non-anchor image, but in the example I tried
they had no noticeable effect on the end result.
Roger Broadie
--
A list of frequently asked questions is available at:
http://wiki.panotools.org/Hugin_FAQ
---
You received this message because you are subscribed to the Google Groups "hugin and other free panoramic software" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to hugin-ptx+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/hugin-ptx/544788BA.7030903%40ogea.freeserve.co.uk.
For more options, visit https://groups.google.com/d/optout.
