[GRASS-user] Peak point extraction from DEM

[Michael Barton]

Mars,

I see that you've received several good replies. I'll add a bit more here.


On 7/2/07 12:00 AM, "Amagine" <amagine at telus.net> wrote:

> Thanks Martin, I'll have a look at terrain analysis in grass.  See
> what I can come up with.
> 
>>> I would like to be able to regulate the following parameters:
>>> 
>>> Minimum height of peak
>> 
>> A topographic "peak" will have an elevation, not a range of
>> elevations. So
>> I'm not sure what you mean.
> 
> I should have elaborated further.
> 
> Minimum height of peak, meaning do not classify any peaks below a set
> minimum threshold : 1000 m for instance.

It looks like by "peak", you mean mountain and by "summit" you mean
topographic peak.

Use r.recode (you could also use r.mapcalc or r.class) to isolate all areas
above 1000m. The recode rule would be...

1000:5000:1000:5000


...by default the rest will go to NULL, giving you a DEM with only places
above 1000m and less than 5000m.

As Maciek, suggested, you could use r.clump to give an identifier to each
area above the cutoff.

output map = peaks (see below)


> 
> 
>>> Minimum drop of surrounding peaks
>> 
>> Not sure what you mean here too.
> 
> Meaning a peak is not extracted if there are surrounding elevations
> within a minimum drop amount
> eg. 50m or 100m (depending on accuracy of DEM file and refinement
> desired of distinct peaks)

I'm not sure if this is needed if you are isolating everything (and only
everthing) above a certain elevation. That is, a mountain is not singled out
if it is surrounded by equally high mountains and no valleys below the
cutoff point.

> 
> 
>>> summit points
>>> optional hierarchy...
>>> 
>> 
>> I guess these are technical mountaineering terms? So again I'm not
>> sure what
>> data you are hoping to extract.
> 
> Summit points.... The highest point on a peak.

These are the topographic peaks from r.param.scale. Use r.reclass to single
out the summits (cells with value of "peak") only.

output map = summits (see below)

Use the map calculator to make a map of summit elevations, using your summit
map and the original DEM.

r.mapcalc 'summit_elev = if(!isnull(summit),DEM,null())

> 
> So a peak encompasses all vertical surface areas within the "Summit"
> point based on the above two categories.

This doesn't make sense to me. Sorry.

> 
> I suppose I define a peak as a categorical area, and a summit as a
> point. Perhaps there are better definitions.

OK. So "peaks" are areas of topography above a certain elevation; "summits"
are topographic peaks within those areas.

> 
> Hierarchy in my mind would be to classify peaks within peaks. Based
> on the initial two parameters (Minimum height of peak, minimum drop
> of surrounding peaks)

Use r.report to list the summits within each peak.

r.report map=peaks,summit_elev units=meters

...will create a report with the elevation of each summit within each peak.

Is this similar to what you need?

Michael

> 
> I suppose I should draw a picture :) probably make more sense.
> 
> Thanks!
> 
> Mars
> 
> On 1-Jul-07, at 10:03 PM, Michael Barton wrote:
> 
>> 
>> 
>> 
>> On 7/1/07 5:46 PM, "Amagine" <amagine at telus.net> wrote:
>> 
>>> Hello, I am trying to extract peak points from a digital elevation
>>> map.
>> 
>> (GUI menu: raster>terrain analysis>terrain parameters)
>> 
>> r.param.scale input=DEM output=feature.map param=feature
>> 
>> This will ID peaks, ridges, passes, channels, pits, and planes
>> 
>> __________________________________________
>> Michael Barton, Professor of Anthropology
>> Director of Graduate Studies
>> School of Human Evolution & Social Change
>> Center for Social Dynamics & Complexity
>> Arizona State University
>> 
>> phone: 480-965-6213
>> fax: 480-965-7671
>> www: http://www.public.asu.edu/~cmbarton
>> 
>> 
> 
> 

__________________________________________
Michael Barton, Professor of Anthropology
Director of Graduate Studies
School of Human Evolution & Social Change
Center for Social Dynamics and Complexity
Arizona State University

phone: 480-965-6213
fax: 480-965-7671
www: http://www.public.asu.edu/~cmbarton