On Wed, 4 Feb 2009, Moreau, Jean-Sébastien wrote:
Hi all,
I did some test with the roadmatcher and wanted to give you some feedback.
Although the tool has advanced features for the conflation of roads
networks, the automatic matching function doesn't behave as expected when
the two datasets have significant differences in terms of segmentation
rules. As some of you might have realised, the geobase network has a lo
When I work with the data in wgs84 I tend to use5D
min segment: 5.0E-7
Standalone distance: 5.0E-4
Match distance: 5.0E-4
Or if I reproject to meters I tend to have a min segment length of 1 m, and
use valuse between 10-100 meters for the other values depending on how built
on the area is and the quality of the OSM data.
So far I have only been using road matcher to detect standalone roads and
only import the standalones. This does means some roads though
more junction than the osm network. The roadmatcher tool seems to have
trouble to find the different pieces of geobase segment that correspond to
an OSM element. Fixing the resulting layer was a tedious task.
It might be possible though to tweak on the different tolerance values in order
to improve the result.
Have some of you got good result with this tool? If so, what values did you use
for the tolerances?
Right now, I'm working on a simpler solution using fme.
Jean-Sebastien Moreau
_______________________________________________
Talk-ca mailing list
[email protected]
http://lists.openstreetmap.org/listinfo/talk-ca
