[prev in list] [next in list] [prev in thread] [next in thread] 

List:       grass-dev
Subject:    Re: [GRASS-dev] r.relief
From:       Paulo van Breugel <p.vanbreugel () gmail ! com>
Date:       2015-04-30 16:08:03
Message-ID: 55425363.6050805 () gmail ! com
[Download RAW message or body]

Thanks for the explanation, and yes, it does makes sense to leave it as 
an option to the user.

Rgds,

Paulo

On 30-04-15 16:44, Helena Mitasova wrote:
> In areas of low relief the integer DEMs have steps (you get alternating flat and \
> steep areas, with steep areas along contours which you may be seeing in your shaded \
> relief map), so when you run r.relief and you get a nice smooth map because the \
> data were converted to float using bilinear interpolation you do not realize that \
> you have the steps (you don't see them on the 2D elevation map). Then you run some \
> modeling or analysis using slope or flow routing running through these flats (which \
> could be quite large) and you get various artificial patterns of the modeling \
> results in these flat areas or you run some statistical analysis you will get \
> serious bias in the histogram see the slide #14 here
> http://courses.ncsu.edu/mea582/common/GIS_anal_lecture/GIS_Anal_Ltopoanal.ppt
> 
> If you convert to float using nearest neighbor you may still get the flats \
> depending on the resolution. Integer DEMs are pretty much a pain to work with if \
> you are doing more than just visualizing the terrain and the steps along 1m \
> contours are quite hard to get rid of in areas of low relief - you pretty much have \
> to reinterpolate the DEM. Integer shaded relief output alerts user that the data \
> are integer and they may have a problem, or if it is in mountains that the integers \
> are not an issue. 
> So what may be a convenience for one user may cause problems for other users, so if \
> we want to conversion it should be a flag (similar situation like the -a flag in \
> r.watershed). 
> Helena
> 
> 
> 
> > On Apr 30, 2015, at 10:00 AM, Paulo van Breugel <p.vanbreugel@gmail.com> wrote:
> > 
> > Out of curiosity, what are those reasons? From a user perspective it might not be \
> > that obvious (well, it wasn't for me at least). If the current implementation is \
> > maintained, perhaps it would be useful to add a note to the manual page? 
> > Paulo
> > 
> > 
> > 
> > On 30-04-15 15:15, Helena Mitasova wrote:
> > > I think that user should do the conversion (in most cases it needs to be \
> > > reinterpolated). I think that the current behavior is right for many reasons,
> > > 
> > > Helena
> > > 
> > > 
> > > > On Apr 30, 2015, at 8:25 AM, Paulo van Breugel <p.vanbreugel@gmail.com> \
> > > > wrote: 
> > > > Hi,
> > > > 
> > > > When running r.relief (in GRASS7.1) on a integer DEM raster, the output is an \
> > > > integer map. Is this intended behavior? Would it be possible to have r.relief \
> > > > convert the layer to double precision / float automatically if the DEM is of \
> > > > CELL type? 
> > > > Paulo
> > > > _______________________________________________
> > > > grass-dev mailing list
> > > > grass-dev@lists.osgeo.org
> > > > http://lists.osgeo.org/mailman/listinfo/grass-dev
> > > Helena Mitasova
> > > Professor at the Department of Marine,
> > > Earth, and Atmospheric Sciences
> > > and Center for Geospatial Analytics
> > > North Carolina State University
> > > Raleigh, NC 27695-8208
> > > hmitaso@ncsu.edu
> > > http://geospatial.ncsu.edu/osgeorel/
> > > "All electronic mail messages in connection with State business which are sent \
> > > to or received by this account are subject to the NC Public Records Law and may \
> > > be disclosed to third parties." 
> Helena Mitasova
> Professor at the Department of Marine,
> Earth, and Atmospheric Sciences
> and Center for Geospatial Analytics
> North Carolina State University
> Raleigh, NC 27695-8208
> hmitaso@ncsu.edu
> http://geospatial.ncsu.edu/osgeorel/
> "All electronic mail messages in connection with State business which are sent to \
> or received by this account are subject to the NC Public Records Law and may be \
> disclosed to third parties." 

_______________________________________________
grass-dev mailing list
grass-dev@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-dev


[prev in list] [next in list] [prev in thread] [next in thread]