[GRASS-user] Re: wierd artifacts in DEM from seamless.usgs.gov

[Helena Mitasova]

Dylan,

it could be the gdalwrap, I have seen this kind of artifacts coming  
from the projection programs,
actually I got something like this when I projected the 10m spearfish  
data that are in the database with r.proj
with the default method (nearest).

So first import the original latlon DEM into GRASS and view it shaded  
or in nviz to check whether
the original is OK. If it does not have the problem (make sure you  
have the resolution set to the imported raster
so that no resampling is done). Then use r.proj to project it into  
UTM, for method - nearest will give you stripes,
but I believe that I got good result with cubic. BTW the tps option  
in gdalwrap should give you good results
too, but there may be some resampling going on that gives you the  
stripes.

Let me know if it worked,

Helena

On May 18, 2006, at 10:04 PM, Dylan Beaudette wrote:

> Howdy,
>
> downloaded a 1/3 arc-second (about 10m res) DEM from the  
> seamless.usgs.gov
> site.
>
> projected the lat/lon raster to UTM via gdalwarp using the -tps option
>
> import into grass, and compute shaded relief... and it looks like  
> crap:
> http://169.237.35.250/~dylan/temp/seamless_dem_artifacts.png
>
> so far the only way that i have been able to get a somewhat decent  
> DEM from
> this file is to :
>
> a) average the heck out of it via r.neighbors size=15 or so
> b) smooth it with r.param.scale size=15 or so
>
> the artifacts appear to be at a regular interval ...
>
> any thoughts one what could be causing this? I have never had this  
> happen
> before, although I have seen numerous scan-line style errors like  
> this in
> USGS DEM products. thoughts of artifact removal bring to mind the  
> recently
> mentioned Albani and Klinkeberg paper [1] and another by Oimoen [2] .
>
> Any thoughts on how to proceed? perhaps a combination of band  
> filters applied
> in the direction and orthagonal to that of the artifacts?
>
> 1.http://scholar.google.com/url?sa=U&q=http://www.geog.ubc.ca/ 
> courses/geog516/notes/Albani_and_Klinkenberg_ASPRS2003.pdf
>
> 2.http://scholar.google.com/url?sa=U&q=http://edna.usgs.gov/Edna/ 
> pubs/artifacts.pdf
>
>
> I was hoping to use this area in the upcoming Where 2.0 talk, and  
> it is
> looking like I might have to choose an area with better elevation  
> data if
> this cannot be resolved. If it can be, then it would be a great  
> slide to add!
>
> Cheers,
>
> -- 
> Dylan Beaudette
> Soils and Biogeochemistry Graduate Group
> University of California at Davis
> 530.754.7341