[postgis-users] Postgis - Line to Point distance
Peter Kukuča
pjr.kukuca at gmail.com
Fri Jul 10 06:58:15 PDT 2009
Hi,
Thank you a million times Pavel. I tried it out and it really works and
gives the distance from the line. It is absolutely great. I didn't think I
would have it running ever.
But, now I have another problem. Look at these two select statements:
select ST_distance_spheroid(pointfromtext('POINT(60 0)', 4326),
ST_line_interpolate_point(linestringfromtext('LINESTRING(0 0,80 0)', 4326),
ST_line_locate_point(linestringfromtext('LINESTRING(0 0,80 0)', 4326),
pointfromtext('POINT(60 1)', 4326))),
'SPHEROID["GRS_1980",6378137,298.257222101]');
returns the same distance (110574.38861168) as
select ST_distance_spheroid(pointfromtext('POINT(0 1)', 4326),
ST_line_interpolate_point(linestringfromtext('LINESTRING(0 0,80 0)', 4326),
ST_line_locate_point(linestringfromtext('LINESTRING(0 0,80 0)', 4326),
pointfromtext('POINT(0 1)', 4326))),
'SPHEROID["GRS_1980",6378137,298.257222101]');
What is done in the example is a calculation of the distance represented by
1 degree of longitude at two different latitudes. While the distance at zero
latitude is correct, the distance at 60° latitude should be exactly half
that. The funny part is, that I get the same result with simple point to
point calculations. And the absolutely best part is, when I swap all
latitudes and longitudes, it works. Now the only reasonable explanation is,
that all of the functions take longitude as their first parameter and
latitude as the second, but isn't this the other way around?
2009/7/10 Pavel Iacovlev <iacovlev.pavel at gmail.com>
> ST_line_locate_point returns the location (float) of the closest point
> on a line to a specified point, read the postgis manual it's all
> written there. After this you use ST_line_interpolate_point to
> transform that location to a point geometry.
>
> 2009/7/10 Peter Kukuča <pjr.kukuca at gmail.com>:
> > Ok, I will be more precise.
> >
> > A line in postgis is represented by a sequence of points (let's call them
> > A,B,C,D...) connected with straight lines, right?
> >
> > Now, for me, the distance between a point (X) and a line in general is
> the
> > smallest distance between any point on the line and the point X. But, by
> any
> > point on the line, I do not mean any of the A,B,C,D... points, but also
> any
> > point on the straight lines between A and B, B and C, C and D....
> >
> > So, the function you suggested - ST_line_locate_point(line, point), does
> it
> > give me one of the A,B,C,D... points or does return really the closest
> > point, even if it has to find it somewhere on the connecting lines? This
> is
> > crucial. If this can really find the closest point, then the rest is
> great,
> > but I doubt that.
> >
> > 2009/7/10 Pavel Iacovlev <iacovlev.pavel at gmail.com>
> >>
> >> The length of the line, you can't calculate a distance between a point
> >> and a line in general. You can calculate a distance between a point
> >> and a point on a line.
> >>
> >> ST_distance_spheroid(point, ST_line_interpolate_point(line,
> >> ST_line_locate_point(line, point)),
> >> 'SPHEROID["GRS_1980",6378137,298.257222101]')
> >>
> >> ST_line_locate_point(line, point) - a point on the line that is most
> >> close to your initial point
> >>
> >> ST_line_interpolate_point - get that point as a point geometry
> >>
> >> ST_distance_spheroid - calculate the length between 2 points
> >>
> >> 2009/7/10 Peter Kukuča <pjr.kukuca at gmail.com>:
> >> > How is your query calculating the distance between a point and a line?
> >> > Are
> >> > you not just calculating the length of the line?
> >> >
> >> > 2009/7/10 Pavel Iacovlev <iacovlev.pavel at gmail.com>
> >> >>
> >> >> Works on lines for me:
> >> >>
> >> >> QUERY:
> >> >> select ST_length_spheroid(st_transform(the_geom, 4326),
> >> >> 'SPHEROID["GRS_1980",6378137,298.257222101]') as ellps_distance,
> >> >> st_distance(the_geom) as normal_distance from fondcart_layers.roads
> >> >> limit 1
> >> >>
> >> >> RESULT:
> >> >> ellps_distance: 9214.65310341668
> >> >> normal_distance: 9214.10811677981
> >> >>
> >> >> 2009/7/10 Peter Kukuča <pjr.kukuca at gmail.com>:
> >> >> > ST_Length_spheroid is perfectly great, but it does not take a line
> as
> >> >> > an
> >> >> > input :-(
> >> >> >
> >> >> > 2009/7/10 Pavel Iacovlev <iacovlev.pavel at gmail.com>
> >> >> >>
> >> >> >> And for geometry length use ST_length_spheroid(geometry,spheroid)
> >> >> >>
> >> >> >> On Fri, Jul 10, 2009 at 11:48 AM, Pavel
> >> >> >> Iacovlev<iacovlev.pavel at gmail.com> wrote:
> >> >> >> > you can take the middle of the line,
> >> >> >> > st_line_interpolate_point(the_geom,
> >> >> >> > 0.5)
> >> >> >> >
> >> >> >> > 2009/7/10 Peter Kukuča <pjr.kukuca at gmail.com>:
> >> >> >> >> Hi Pedro,
> >> >> >> >>
> >> >> >> >> thank you for your reply.
> >> >> >> >>
> >> >> >> >> I cannot use pointn, the line can be up to several hundred
> >> >> >> >> kilometers
> >> >> >> >> long
> >> >> >> >> and consits of up to 800 points, so this would be too rough.
> >> >> >> >>
> >> >> >> >> Regardinf the decond suggestion, that is exactly what I am
> doing
> >> >> >> >> now,
> >> >> >> >> but I
> >> >> >> >> don't know, which projection (srid) to transform the wgs84 data
> >> >> >> >> into.
> >> >> >> >> It
> >> >> >> >> should work on all of the earth's surface, at least in all of
> >> >> >> >> europe
> >> >> >> >> for a
> >> >> >> >> start. And from what I found, the projections are usually tied
> to
> >> >> >> >> a
> >> >> >> >> small
> >> >> >> >> territory. I was thinking about mercator projection, but there
> is
> >> >> >> >> a
> >> >> >> >> zillion
> >> >> >> >> of them in postgis. And I found a post, that the one with srid
> >> >> >> >> 900913
> >> >> >> >> is
> >> >> >> >> working great, but I don't have that one in my postgis.
> >> >> >> >>
> >> >> >> >> 2009/7/10 Pedro Doria Meunier <pdoria at netmadeira.com>
> >> >> >> >>>
> >> >> >> >>> -----BEGIN PGP SIGNED MESSAGE-----
> >> >> >> >>> Hash: SHA1
> >> >> >> >>>
> >> >> >> >>> Hi Peter
> >> >> >> >>>
> >> >> >> >>> You could use the pointn(geometry, n) for the first point of
> the
> >> >> >> >>> linestring as a first approach.
> >> >> >> >>> As far as meters are concerned here's a little example of
> >> >> >> >>> transforming
> >> >> >> >>> the geometry to the desired projected system:
> >> >> >> >>>
> >> >> >> >>> select distance(transform(u.coordinates,srid),
> >> >> >> >>> transform(p.geometry,srid)) AS thedistance
> >> >> >> >>>
> >> >> >> >>> HTH,
> >> >> >> >>>
> >> >> >> >>> Pedro Doria Meunier
> >> >> >> >>> GSM: +351 96 17 20 188
> >> >> >> >>> Skype: pdoriam
> >> >> >> >>>
> >> >> >> >>>
> >> >> >> >>>
> >> >> >> >>>
> >> >> >> >>> Peter Kukuča wrote:
> >> >> >> >>> > Dear sir,
> >> >> >> >>> >
> >> >> >> >>> > I am not sure if I am writing to the correct email address.
> I
> >> >> >> >>> > have
> >> >> >> >>> > found a post on the postgis-users formu from *Rich Gibson
> *and
> >> >> >> >>> > this
> >> >> >> >>> > address was next to it.
> >> >> >> >>> >
> >> >> >> >>> > I am using postgis for a while now and it is great. It
> solved
> >> >> >> >>> > a
> >> >> >> >>> > lot
> >> >> >> >>> > of problems for me.
> >> >> >> >>> >
> >> >> >> >>> > However, I am now facing a problem I cannot solve. I
> searched
> >> >> >> >>> > for
> >> >> >> >>> > almost two days now, but I still cannot find an answer. Here
> >> >> >> >>> > is
> >> >> >> >>> > my
> >> >> >> >>> > problem:
> >> >> >> >>> >
> >> >> >> >>> > I have a database of linestrings in wgs84 projection and
> >> >> >> >>> > points
> >> >> >> >>> > in
> >> >> >> >>> > wgs84 projection. I would like to determine the distance
> >> >> >> >>> > between
> >> >> >> >>> > a
> >> >> >> >>> > line and a point. The ST_distance function does this very
> >> >> >> >>> > well,
> >> >> >> >>> > but
> >> >> >> >>> > it does not take the wgs84 projection into account and also,
> >> >> >> >>> > the
> >> >> >> >>> > result is in degrees. On the other hand, the
> >> >> >> >>> > ST_Distance_Spheroid
> >> >> >> >>> > does take the wgs84 into accound and does give the result in
> >> >> >> >>> > meters,
> >> >> >> >>> > but it does not accept a linestring as an input parameter.
> >> >> >> >>> >
> >> >> >> >>> > Is there any way i can the distance between a line and a
> point
> >> >> >> >>> > in
> >> >> >> >>> > meters from the wgs84 projected input? I do not need grat
> >> >> >> >>> > accuracy.
> >> >> >> >>> > +-5 meters is still good enough. Thank you for your tips.
> >> >> >> >>> >
> >> >> >> >>> > --
> >> >> >> >>> > S pozdravom (regards)
> >> >> >> >>> > Ing. Peter Kukuča
> >> >> >> >>> >
> >> >> >> >>> >
> >> >> >> >>> >
> >> >> >> >>> >
> >> >> >> >>> >
> ----------------------------------------------------------------------
> >> >> >> >>> >
> >> >> >> >>> > _______________________________________________
> >> >> >> >>> > postgis-users mailing list
> >> >> >> >>> > postgis-users at postgis.refractions.net
> >> >> >> >>> >
> http://postgis.refractions.net/mailman/listinfo/postgis-users
> >> >> >> >>> -----BEGIN PGP SIGNATURE-----
> >> >> >> >>> Version: GnuPG v1.4.7 (GNU/Linux)
> >> >> >> >>> Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org
> >> >> >> >>>
> >> >> >> >>>
> iD8DBQFKVv052FH5GXCfxAsRAqS1AJ0dfphFX3wrMuo+FdXVRTgmUzHRUACfUmvH
> >> >> >> >>> 146yGHfxoVmlnt9a91rpWmE=
> >> >> >> >>> =G70f
> >> >> >> >>> -----END PGP SIGNATURE-----
> >> >> >> >>>
> >> >> >> >>> _______________________________________________
> >> >> >> >>> postgis-users mailing list
> >> >> >> >>> postgis-users at postgis.refractions.net
> >> >> >> >>> http://postgis.refractions.net/mailman/listinfo/postgis-users
> >> >> >> >>
> >> >> >> >>
> >> >> >> >>
> >> >> >> >> --
> >> >> >> >> S pozdravom (regards)
> >> >> >> >> Ing. Peter Kukuča
> >> >> >> >>
> >> >> >> >> _______________________________________________
> >> >> >> >> postgis-users mailing list
> >> >> >> >> postgis-users at postgis.refractions.net
> >> >> >> >> http://postgis.refractions.net/mailman/listinfo/postgis-users
> >> >> >> >>
> >> >> >> >>
> >> >> >> >
> >> >> >> >
> >> >> >> >
> >> >> >> > --
> >> >> >> > http://iap.md, The future is open
> >> >> >> >
> >> >> >>
> >> >> >>
> >> >> >>
> >> >> >> --
> >> >> >> http://iap.md, The future is open
> >> >> >> _______________________________________________
> >> >> >> postgis-users mailing list
> >> >> >> postgis-users at postgis.refractions.net
> >> >> >> http://postgis.refractions.net/mailman/listinfo/postgis-users
> >> >> >
> >> >> >
> >> >> >
> >> >> > --
> >> >> > S pozdravom (regards)
> >> >> > Ing. Peter Kukuča
> >> >> >
> >> >> > _______________________________________________
> >> >> > postgis-users mailing list
> >> >> > postgis-users at postgis.refractions.net
> >> >> > http://postgis.refractions.net/mailman/listinfo/postgis-users
> >> >> >
> >> >> >
> >> >>
> >> >>
> >> >>
> >> >> --
> >> >> http://iap.md, The future is open
> >> >> _______________________________________________
> >> >> postgis-users mailing list
> >> >> postgis-users at postgis.refractions.net
> >> >> http://postgis.refractions.net/mailman/listinfo/postgis-users
> >> >
> >> >
> >> >
> >> > --
> >> > S pozdravom (regards)
> >> > Ing. Peter Kukuča
> >> >
> >> > _______________________________________________
> >> > postgis-users mailing list
> >> > postgis-users at postgis.refractions.net
> >> > http://postgis.refractions.net/mailman/listinfo/postgis-users
> >> >
> >> >
> >>
> >>
> >>
> >> --
> >> http://iap.md, The future is open
> >> _______________________________________________
> >> postgis-users mailing list
> >> postgis-users at postgis.refractions.net
> >> http://postgis.refractions.net/mailman/listinfo/postgis-users
> >
> >
> >
> > --
> > S pozdravom (regards)
> > Ing. Peter Kukuča
> >
> > _______________________________________________
> > postgis-users mailing list
> > postgis-users at postgis.refractions.net
> > http://postgis.refractions.net/mailman/listinfo/postgis-users
> >
> >
>
>
>
> --
> http://iap.md, The future is open
> _______________________________________________
> postgis-users mailing list
> postgis-users at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-users
>
--
S pozdravom (regards)
Ing. Peter Kukuča
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/postgis-users/attachments/20090710/189fba6c/attachment.html>
More information about the postgis-users
mailing list