Re: [postgis-users] Evenly distributing a point set
Replied at http://gis.stackexchange.com/questions/131854/spacing-a-set-of-pointswith some Postgis options... Brent Wood From: toni hernández t...@sigte.udg.edu To: postgis-users@lists.osgeo.org Sent: Wednesday, January 28, 2015 12:05 AM Subject: Re: [postgis-users] Evenly distributing a point set Maybe not on a point. I was thinking with linestring On 27/01/2015 11:46, Dave Barter wrote: #yiv5480649879 body{font-family:Helvetica, Arial;font-size:13px;} Maybe you can also try to simplify the geometry using X as the tolerance. http://postgis.net/docs/manual-1.4/ST_Simplify.html On 25/01/2015 12:54, Dave Barter wrote: KNN distance Will that work on a point layer? I have added a diagram and more detail here http://gis.stackexchange.com/questions/131854/spacing-a-set-of-points ___ postgis-users mailing list postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users ___ postgis-users mailing list postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users ___ postgis-users mailing list postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users
Re: [postgis-users] Evenly distributing a point set
Maybe you can also try to simplify the geometry using X as the tolerance. http://postgis.net/docs/manual-1.4/ST_Simplify.html On 25/01/2015 12:54, Dave Barter wrote: KNN distance Will that work on a point layer? I have added a diagram and more detail here http://gis.stackexchange.com/questions/131854/spacing-a-set-of-points___ postgis-users mailing list postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users
Re: [postgis-users] Evenly distributing a point set
Maybe not on a point. I was thinking with linestring On 27/01/2015 11:46, Dave Barter wrote: Maybe you can also try to simplify the geometry using X as the tolerance. http://postgis.net/docs/manual-1.4/ST_Simplify.html On 25/01/2015 12:54, Dave Barter wrote: KNN distance Will that work on a point layer? I have added a diagram and more detail here http://gis.stackexchange.com/questions/131854/spacing-a-set-of-points ___ postgis-users mailing list postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users ___ postgis-users mailing list postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users
Re: [postgis-users] Evenly distributing a point set
Maybe you can also try to simplify the geometry using X as the tolerance. http://postgis.net/docs/manual-1.4/ST_Simplify.html On 25/01/2015 12:54, Dave Barter wrote: KNN distance ___ postgis-users mailing list postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users
Re: [postgis-users] Evenly distributing a point set
Hey, I used Nicolas's strategy and it works. You could slightly improving it by some preprocessing on the linestring itself (smoothing or simplifying, depending on your data) Cheers, Rémi-C 2015-01-25 22:32 GMT+01:00 Nicolas Ribot nicolas.ri...@gmail.com: Hello, You could build the linestring from points (makeLine(geom order by time)) then , with linear referencing (st_lineInterpolatePoint, st_linelocatePoint), you can divide your linestring by the distance you want, creating new points. Then, if you need to keep time information, you could interpolate time from initial segment where a new point lies in. Nicolas On 25 January 2015 at 12:54, Dave Barter d...@phased.co.uk wrote: If I had a table of points which are irregularly distributed and wanted to evenly distribute them what would be the best strategy and query? These points were created from a GPX file along a road journey so they crudely map to a line string. I was thinking compute the KNN distance for each point and remove those that have a KNN a threshold (say 100 metres) Is that the best way to go about it? -- Dave Barter Web: http://www.phased.co.uk Email:d...@phased.co.uk Twitter:@Citizenfishy ___ postgis-users mailing list postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users ___ postgis-users mailing list postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users ___ postgis-users mailing list postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users
Re: [postgis-users] Evenly distributing a point set
Hey, it is unclear what you want and how your data looks like. For better answer it would be good to have better info. Maybe a screen of your data, some facts, for what usage (admissible error, expected quality ..) etc etc. What you are trying to do (reconstruct a network from sampled pieces) is in general a complicated problem. It may be simpler depending on your data and hypothesis. Cheers, Rémi-C 2015-01-26 9:28 GMT+01:00 Dave Barter d...@phased.co.uk: These points were created from a GPX file along a road journey so they crudely map to a line string Map matching is the term often used to describe the process of assigning noisy Gpx points to underlying road network. My problem is a bit more involved. The GPX points represent the national cycle network in the UK. These points cover roads/tracks/paths and all sorts and as such I cannot really snap to map. Creating line strings from the points will possibly cause errors as well as I think there will be a lot of cases where the wrong points are joined. I think the safest method is to remove densely distributed points from the current set until they all have a KNN value of X. This will not create a completely even distribution but will ensure that no points are created in the wrong place which may be the case with line interpolation. Can anyone help me with a query to do this? ___ postgis-users mailing list postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users ___ postgis-users mailing list postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users
Re: [postgis-users] Evenly distributing a point set
These points were created from a GPX file along a road journey so they crudely map to a line string Map matching is the term often used to describe the process of assigning noisy Gpx points to underlying road network. My problem is a bit more involved. The GPX points represent the national cycle network in the UK. These points cover roads/tracks/paths and all sorts and as such I cannot really snap to map. Creating line strings from the points will possibly cause errors as well as I think there will be a lot of cases where the wrong points are joined. I think the safest method is to remove densely distributed points from the current set until they all have a KNN value of X. This will not create a completely even distribution but will ensure that no points are created in the wrong place which may be the case with line interpolation. Can anyone help me with a query to do this?___ postgis-users mailing list postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users
Re: [postgis-users] Evenly distributing a point set
Hello, You could build the linestring from points (makeLine(geom order by time)) then , with linear referencing (st_lineInterpolatePoint, st_linelocatePoint), you can divide your linestring by the distance you want, creating new points. Then, if you need to keep time information, you could interpolate time from initial segment where a new point lies in. Nicolas On 25 January 2015 at 12:54, Dave Barter d...@phased.co.uk wrote: If I had a table of points which are irregularly distributed and wanted to evenly distribute them what would be the best strategy and query? These points were created from a GPX file along a road journey so they crudely map to a line string. I was thinking compute the KNN distance for each point and remove those that have a KNN a threshold (say 100 metres) Is that the best way to go about it? -- Dave Barter Web: http://www.phased.co.uk Email:d...@phased.co.uk Twitter:@Citizenfishy ___ postgis-users mailing list postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users ___ postgis-users mailing list postgis-users@lists.osgeo.org http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users