[osg-users] Image Resolution in VPB

Martins Innus minnus at ccr.buffalo.edu
Thu Mar 19 12:27:23 PDT 2009

	Thanks. Using -e with appropriate options did exactly what I wanted, 
except for a slight error on the edges of the created images.  If i 
overlay the created image tiles from the highest resolution level on top 
of the original source imagery, they are pixel to pixel exact except for 
a 1 pixel strip all the way around the image.
	It appears that the edge values are an average of what the pixel value 
actually should be and the pixel directly adjacent to it but what should 
be the next tile.
	If you layer the attached good and bad images on top of each other in 
an image viewer and toggle back and forth you can see the pixels changing.
	The good image is just the original source imagery zoomed in.  The bad 
image shows the vpb generated tile overlayed  on the left side.

	I saw the "corner equalization bug" posts, but this seems like a 
different issue.

This is under Redhat Enterpise Linux, x86_64, with OSG 2.8 and VPB svn 
of a couple days ago.


Robert Osfield wrote:
> Hi Martins,
> You could try to use the -e option.  It takes lat/longs as input.  I 
> don't know if it'll pad though, as I suspect it won't unless the at 
> least some of the input data covers the region.   If you want to force 
> the resolution to be the same then just use the image and height res 
> option.  I can't recall what they off the top of my head, run osgdem 
> --help to list them all.
> Robert.
> On Mon, Mar 16, 2009 at 6:12 PM, Martins Innus <minnus at ccr.buffalo.edu 
> <mailto:minnus at ccr.buffalo.edu>> wrote:
>     Hello,
>            I'm using VPB to generate a terrain database with 2ft
>     resolution elevation data and 1 ft resolution imagery. What I'm
>     trying to do is make the imagery in the highest resolution tiles
>     have the exact same resolution as the source imagery and also still
>     be a power of 2 for the image tiles.
>            I've verified that if I take a small input area where the
>     source imagery dimensions are a power of 2, I get what i want for
>     the final model.
>            It seems that if I could tell VPB to "pad out" the extents of
>     my input data to a power of 2 that should work as well.  I don't
>     care if its black or garbage or whatever.
>            Would the "-e" option do what I want if I specify extents
>     larger than the input data?  I'm about to give that a try but it
>     takes about three days for this job to run to completion, so I
>     figured I would ask if anyone has tried to do the same thing.
>     Thanks for any insight.
>     Martins
>     _______________________________________________
>     osg-users mailing list
>     osg-users at lists.openscenegraph.org
>     <mailto:osg-users at lists.openscenegraph.org>
>     http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
> ------------------------------------------------------------------------
> _______________________________________________
> osg-users mailing list
> osg-users at lists.openscenegraph.org
> http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
-------------- next part --------------
A non-text attachment was scrubbed...
Name: good_tex.jpg
Type: image/jpeg
Size: 14269 bytes
Desc: not available
URL: <http://lists.openscenegraph.org/pipermail/osg-users-openscenegraph.org/attachments/20090319/f219edfe/attachment-0006.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: bad_tex.jpg
Type: image/jpeg
Size: 14316 bytes
Desc: not available
URL: <http://lists.openscenegraph.org/pipermail/osg-users-openscenegraph.org/attachments/20090319/f219edfe/attachment-0007.jpg>

More information about the osg-users mailing list