s.voronoi

Dave Gerdes dpgerdes at zorro.cecer.army.mil
Tue Apr 14 15:29:07 EDT 1992


> To: grassp-list at amber.cecer.army.mil
> Subject: Re: s.voronoi
> Date: Mon, 13 Apr 1992 15:07:32 +1000
> From: murray bent <algebras at cc.uq.oz.au>
> Sender: algebras at cc.uq.oz.au
> 
> > ... I would think that there are other ways around the
> > problem, for example never going to polygons in the first place.  How
> > about building a network instead of a bunch of polygons.  You should never
> > have to re-build any given segment more than once anyway.  if you are, 
> > I would think you are doing twice the necessary work.  
>  
>  	Could you please spell out the correct way to create a network in
> GRASS ? Often , I do not create areas properly if I don't resort to poly-
> gons. Each area appears to need a unique node,  as a rule. If I permute
> the points in one of the area lines there's no problem.  
> 	That is,   A 4  0 0  0 1   1 1  0 0      and
>  		   A 4  0 0  0 1  -1 -1    0 0   
> will not form 2 areas , but the permutation 
>  		   A 4  0 1  -1 -1    0 0  0 1  will work, but this is still
> a polygon representation duplicating line 00 01.
> 
> 	Where is the correct protocol for a network representation ?
> 
> 					Murray
> 

  I am not totally sure what you are saying above, but there might be
one mis-understanding here.  The Area entries in the dig and dig_ascii
files do not have to represent entire closed polygons.  They can represent
line (or arc, polyline) sections of 2 or more vertices which when combined 
with all the other line sections in the file create closed polygons.  for
example, to represent a square, The following would be acceptable:

A 2
   0 0
   0 1
A 2
   0 1
   1 1
A 3
   1 1
   1 0
   0 0

Note those lines can be in any order in the file.   The program v.support
then goes through all that and builds polygons by finding common endpoints
of the line segments.  

So what I was recommending was that each line segment of 2 vertices be 
placed in the file seperately.  Now GRASS does not have a 'network' format
other than the internal topologic structure built by v.support is in effect
a type of network.  I just thought if you are having trouble with duplicate
lines using polygons, that another type of data representation, such
as some type of network notation, might make it simpler to identify 
paths already taken.

If you have more questions, lets drop this into Email instead of bothering 
the net.

-- 


  Dave Gerdes
  US Army Construction Engineering Research Lab
  Spatial Analysis & Systems Team
  dpgerdes at cerl.cecer.army.mil
  (217) 352-6511 x591



More information about the grass-dev mailing list