I was thinking more of the woodland polygons, and the linear features. They are a lot harder to run a complex feature simplification algorithm on than the building outlines. My test area is Warton Crag in north Lancashire. There is a lot of stuff there that looks like it has been hand traced from OS open street view rasters, which I suspect have been translated to WGS84 without using the OSTN02 data.
Roger
The shapes themselves aren’t particularly accurate, if you look at say building outlines and compare to Bing. It shouldn’t take long to find a non-rectangular building on Bing which has been approximated to rectangular in OS Open Map Local
Ed
From: [email protected] [mailto:[email protected]]
Sent: 06 June 2016 14:45
To: [email protected]
Subject: [Talk-GB] OS open map local polygon accuracy
Hi,
I have been looking at the OS openmap local vector dataset. I noticed that the coordinates in there are centimetre level accuracy. I am speculating how the OS made this dataset a "nominal viewing scale" of 1:10000. Scales are somewhat irrelevant to vector data. Have they degraded the geometry points by thinning or averaging, or is the data still at survey level accuracy? I have some old (paid for) OS master map data. It would be interesting to compare the polygons in there with the the openmap local ones. But before I search my loft for the disc, has anyone already done this?
Roger
_______________________________________________ Talk-GB mailing list [email protected] https://lists.openstreetmap.org/listinfo/talk-gb
