[GRASSLIST:690] Re: height contour lines to raster

[Maciek Sieczka]

On Wed, 12 Apr 2006 15:15:53 +0200
Jachym Cepicky <jachym.cepicky at centrum.cz> wrote:

> v.surf.rst (?)

Worth of trying, but please expect problems (segmentation, artifficial
hollows, bulges). The more distant and clustered your input contour
lines are, the more of those artifacts. RST is great for fairly regular
input data, but problematic for most real elavation contour line
datasets. Of course, you might perfectly be happy with the result,
depending on what you expect of your DEM.

r.surf.contour is another way to go but also limited - it gets freaky
with highly clustered data, requires contour lines to be perfectly
continous, no elevation spots between the contour lines. Also it
creates integer DEM only, but this can be workedaround by
multiplying your input and dividing the output by the same number, say
1000.

For DEM interpolation I've personally been using nnbathy natural
neighbor interpolator, and I'm happy with it as it doesn't have all the
limitations of the two above (not mentioning IDW, no good for DEM from
disperse, clustered contour lines+points). Yet it is quite cumbersone to
make it use Grass data and import it's output back to Grass. If you'd
like to give a try, look at my rough how-to here:
http://grass.itc.it/pipermail/grass5/2006-April/022171.html

It would be great if natural neighbor interpolator was implemented in
Grass IMHO.

Any ideas how to improve my nnbathy-Grass "interface" are wellcome.

Cheers
Maciek

--------------------
W polskim Internecie s? setki milion?w stron. My przekazujemy Tobie tylko najlepsze z nich!
http://katalog.panoramainternetu.pl/