I think it's worth adjusting the 'centre' line. I haven't got to the point with my experiments <http://sk53-osm.blogspot.co.uk/2016/10/skeletons-in-water.html> with straight skeletons & approximate medial lines to be comfortable that there is a relatively easy way to derive decent centre-lines. At least for know my problems are of two types:
- Even with the PostGIS AML function one ends up with a multilinestring with lots of dangles. So one option is to recursively prune the structure. - An alternative way to find a path through the AML (or Straight Skeleton) is to route through it. I've done some little experiments with pg_routing which work fine, but havent thought about how such things scale. If I've missed some well-known algorithms for these purposes do let me know. Jerry On 27 March 2017 at 00:30, Richard Welty <[email protected]> wrote: > been editing a bunch in the area around Albany & Troy New York. > there's a linear waterway feature which is supposed to represent > the Hudson River, in addition to the polygons that somewhat accurately > represent the river. the waterway feature isn't terribly accurate, in fact > it's rather ugly looking on the map. here is one location where you > see the straight line off to the left of the semi-accurate polygons: > > http://www.openhistoricalmap.org/go/ZdrHMUa~--?layers=H > > so what value is it bringing? delete or adjust to the rough > centerline of the river? > > richard > > -- > [email protected] > Averill Park Networking - GIS & IT Consulting > OpenStreetMap - PostgreSQL - Linux > Java - Web Applications - Search > > > _______________________________________________ > Historic mailing list > [email protected] > https://lists.openstreetmap.org/listinfo/historic >
_______________________________________________ Historic mailing list [email protected] https://lists.openstreetmap.org/listinfo/historic
