[GRASS5] Re: [GRASSLIST:1477] small raster regions elimination

[Markus Neteler]

Hi Rado

[cc to grass5]

On Wed, Feb 21, 2001 at 03:12:41AM -0600, Rado Bonk wrote:
> Hi Markus
> r.reclass.are does the job, but after eliminating specified regions, there
> are wholes (NO DATA )in the geographic region raster. The regions smaller
> than specified value have to be appended to neighboring regions, 
> based on some reasonable algorithm (for example, based on the
> length of its common boundary > they will be attached to region 
> with longest common boundary). Deleting the regions doesn't solve the
> problem.

So you need a "filling hole" algorithm, what you wrote about inwards
buffering. Now I understand. This is what I am dreaming of as well :-)

Justin: Once we have been talking about this how to fill NULL holes in
a raster map. It seems we need it now.

 Markus

> 
> Rado Bonk
> Geography and Geology
> UNO, Omaha, NE 68182-0199
> rado at cosmos.unomaha.edu
> http://rbonk.host.sk
> 
> On Wed, 21 Feb 2001, Markus Neteler wrote:
> 
> > Hi Rado,
> > 
> > On Wed, Feb 21, 2001 at 02:22:24AM -0600, Rado Bonk wrote:
> > > Hi GRASS users, 
> > > After mulptiply overlay operations, there are several
> > > (hundreds) raster regions with the area too small for analysis. I would
> > > like to eliminate these regions and append them to the adjacent regions
> > > based on the length of its common boundary (that would be the best).
> > > r.reclass.area script does a nice job with elimination of these regions by
> > > assigning the area of the smallest region, and deleting all regions with
> > > area smaller than assigned value. (Markus, there is an error in
> > > line 9, missing || operand, see correct line at the end of the message).
> > Thanks for notification. I had fixed it recently and also another script 
> > bug.
> >  You can catch the latest from (web)CVS.
> > 
> > > The task can be solved by exporting raster regions into vector, and than
> > > using vector algebra in some of the desktop mapping systems. However, I
> > > would like to avoid this, and I would like to use raster data
> > > format. Maybe some filter would help.
> > > 
> > > There was similar question a time ago concerning "inward
> > > buffering" operation with no response.
> > 
> > Sorry, I don't get the problem. If r.reclass.area does the job, what's left?
> > 
> > Markus
> > 
> 

-- 
Markus Neteler *  University of Hannover
Institute of Physical Geography and Landscape Ecology
Schneiderberg 50 * D-30167 Hannover * Germany
Tel: ++49-(0)511-762-4494  Fax: -3984

---------------------------------------- 
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo at geog.uni-hannover.de with
subject 'unsubscribe grass5'