At 2010-05-20 02:44, Jean-Guilhem Cailton wrote:
Le 20/05/2010 06:29, John Smith a écrit : > Here's a sample generated from NASA site log files: > > <node id='-1' visible='true' lat='44.463944444444' lon='26.125738888889'> > <tag k='fixme' v='not_reviewed' /> > <tag k='man_made' v='monitoring_station' /> > <tag k='monitoring:gps' v='yes' /> > <tag k='monitoring:glonass' v='yes' /> > <tag k='iers_domes_number' v='11401M001' /> > <tag k='antenna' v='LEIAT504GG LEIS' /> > <tag k='receiver' v='LEICA GRX1200GGPRO' /> > <tag k='name' v='Bucuresti / Romania' /> > <tag k='ele' v='143.2' /> > <tag k='source:url' > v='http://igscb.jpl.nasa.gov/igscb/station/log/bucu_20100503.log' /> > </node> > > Just to make things more interesting, the lat/lon given for some/all > sites are in country or region specific datums, eg Australian > locations use GDA94, but the site log file can't automatically be > parsed for which datum is used.
And it doesn't even always specify them :(
Compared with the monitoring sites with other purposes (nice aircraft noise tracking site by the way), I thought that at least for the positioning systems monitoring sites, precise lat/lon should be available. Looking at the Toulouse log file (and Bucarest's as well), it seemed to use ITRF reference system (International Terrestrial Reference Frame - http://itrf.ensg.ign.fr/), apparently fairly close to WGS84 for OSM purposes (ftp://itrf.ensg.ign.fr/pub/itrf/WGS84.TXT).
The ITRF values do appear to be precisely what is being measured. However, the "Latitude" and "Longitude" below them are not always correctly transformed to WGS84. I spot-checked a few in my area, finding one that was correct, and some that were off by as much as 100m. I used GeoTrans3 and the WGS84 datum for the transformation. I looked up the Wikipedia source and was satisfied that this is supposedly sufficient to get within 10 cm.
The lat/lon for BUCU is correct under these circumstances, but that for TOUL is not. I get 43 33 38.5307, 1 28 51.2087, 211.655 - ~9 meters away from the spec'd lat/lon, which seems closer to what appears to be the correct structure in the Bing imagery - the square pad near the edge of the building about 2.7m SSW of those coords.
I'd also add ref=* for the 4-char identifier (e.g. BUCU) and start_date from Item 1 "Date Installed".
Note that at least one of these IGSC stations (LEEP) was previously imported by someone: http://www.openstreetmap.org/browse/node/740570067
-- Alan Mintz <alan_mintz+...@earthlink.net> _______________________________________________ Tagging mailing list Tagging@openstreetmap.org http://lists.openstreetmap.org/listinfo/tagging
