[GRASSLIST:7765] Re: Hutchinson's Adaptive Alogrithm for sound DEMs?

[Markus Neteler]

Here a FWD from Helena Mitasova:

On Tue, Aug 02, 2005 at 07:01:33PM -0400, Helena Mitasova wrote:
> regarding watershed analysis, I had troubles getting watersheds using 
> r.water.outlet but I had excellent results for stream extraction and
> watershed boundaries using r.watershed
> for various types of DEMs including those interpolated using 
> s.surf.rst/v.surf.rst.
> 
> Mike's method is very good and I highly recommend it .
> Spatially variable smoothing is already available in 
> s.surf.rst/v.surf.rst, we just did
> not integrate its automated estimation based on slope because this is 
> very specific for contour data and there are many types of data
> where adjusting smoothing as a function of
> slope is not the best approach and can lead to the loss of features.
> Original topogrid had serious problems with waves along contours and 
> this was an excellent
>  solution for the problem. From the email, it is not possible to tell 
> what quality problems
> the user has and whether they can be fixed by tuning the parameters. 
> For example I noticed
> that v.surf.rst had very low npmin default = 200 and a segmentation 
> update that we did for
> s.surf.rst was not implemented. This should be fixed in 6.1
> 
> We are currently testing different interpolation methods including 
> topogrid (or whatever
> its current name is) and v.surf.rst along with other methods and we 
> plan to implement modifications/improvements that we find beneficial.
> 
> I hope that this helps
> 
> Helena
> 
> 
> >On Mon, Aug 01, 2005 at 08:06:21PM +0000, paallen at attglobal.net wrote:
> >>Hi all,
> >>
> >>I am trying to create a DEM from Topographic
> >>contours.  I have been using regularized spline
> >>with tension (v.surf.rst) and I am just not
> >>content with the quality.  My final objective is
> >>to produce watersheds for sample locations using
> >>"r.water.outlet".  So I am looking for
> >>alternatives and as aways preferabley something in
> >>GRASS or at least GNU.
> >>
> >>I was wondering if M.F. Hutchinson's alogrithm for
> >>generating DEMs from his paper on the net "A
> >>Locally Adaptive Approach to the Interprolation of
> >>Digital Elevation Models"
> >>(http://www1.gsi.go.jp/geowww/globalmap-gsi/gtopo30/papers/local.html)
> >>has been written for GRASS or any other GNU program?.
> >>
> >>It appears to be the algorithm that the CGIAR used
> >>to re-grid the SRTM 90m DEM.
> >>
> >>
> >>Thanks,
> >>
> >>Phillip J. Allen
> >>Consulting Geochemist
> >>Lima Peru
> >>e-mail: paallen at attglobal.net
> >>
> >>
> Helena Mitasova
> Dept. of Marine, Earth and Atm. Sciences
> 1125 Jordan Hall, NCSU Box 8208,
> Raleigh NC 27695
> http://skagit.meas.ncsu.edu/~helena/