[postgis-devel] Re: Rounding off coordinates (from Postgis-users)

strk strk at keybit.net
Fri Jul 2 02:18:53 PDT 2004


On Fri, Jul 02, 2004 at 10:34:24AM +0200, Markus Schaber wrote:
> Hi, Strk,
> 
> On Thu, 1 Jul 2004 17:41:22 +0200
> strk <strk at keybit.net> wrote:
> 
> > What I implemented did not remove points with 'too-near' checks.
> > I tough you were doing something similar. RoundingOff primary
> > purpose should be Rounding, not points reduction.
> > What I did is removing points which made really no-sense keeping,
> > which is consecutive points having the same coordinates and 
> > multiple back-and-forth between two points.
> 
> Okay, this makes sense, you're right - this should not create self
> intersections. I was biased by the simplify() function which can create
> invalid, self-intersecting geometries (or at least did when I tested
> last.)

The simplify() function is a DP implementation. 
It's main purpose is removing points...

There are published researches about a non-currupting
simplification algorithm but no one is simple...

Still, you should consider whole-map discrepancies (have you
tried simplifying a world map ?) as non-self-intersecting
polygons will still loose topoloogical relation with neighbours.

It would be interesting to provide a map simplification function.
That would probably be better implemented in GEOS which has indexing
and noding already implemented...

--strk;

> 
> Thanks,
> Markus
> 
> -- 
> markus schaber | dipl. informatiker
> logi-track ag | rennweg 14-16 | ch 8001 z├╝rich
> phone +41-43-888 62 52 | fax +41-43-888 62 53
> mailto:schabios at logi-track.com | www.logi-track.com



More information about the postgis-devel mailing list