As I understand it, enblend selects the seam width based on its
determination of feature size -- the smaller the local feature size
(the more detail there is), the narrower the seam.

The problem here is that when the feature size is anisotropic -- for
example, the local features consist of long horizontal lines --
enblend doesn't pick the seam width very well.  The attached image
fragment demonstrates this.  The local features are long, more or less
featureless horizontal lines, so there's a lot of detail in the
vertical direction but almost none in the horizontal direction.

If a seam crosses these lines, enblend selects a narrow seam because
there's a fair amount of local detail.  The problem is that there
isn't really much detail across the seam boundary.  In this case, the
horizontal lines were rows of waves in the far distance, so they moved
between shots (there may also have been focus differences playing a
part).

I think that enblend should primarily at the detail perpendicular to
the seam and ignore detail parallel to the seam.  In this case,
there's a fair amount of detail in the vertical direction, parallel to
the seam, but almost nothing perpendicular to it.

-- 
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

<<enblend-issue.tif>>

Reply via email to