On Wed, 2013-11-20 at 10:32 +1100, mick wrote: > To gain some confidence in what I've collected I plan to drive as many roads > in my area of interest (Glen Innes, NSW) several times with my two GPS units > recording and find an acceptable level of coincidence before I add anything > to OSM.
Mick, your GPS should give you some 'quality' measure along with its positional information. Look in the XMl, you will see it under HDOP. The basic GPS system (end to end) could deliver answers within 5 to 6 metres. But there are also 'geometry issues associated with aerial and satelite position that multiplies that error. HDOP is a measure of this geometry problem and you multiply it by the by the systems intrinsic errors. At a hdop of 10 and a satellite / receiver intrinsic error of, say, 6 meters, we have an uncertainty of 60 meters. Thats getting a bit high for most use. In some scripts distributed with FoxtrotGPS to clean up tracks for OSM, I set an HDOP limit of 8, points greater than that are rejected. You could, perhaps, apply the same rule to your data. Re-drive the ones that have too many data points greater than 8 HDOP. Now, I have greatly simplified the subject, you can be almost certain someone with a lot more knowledge of DOP and GPS accuracy will clarify it for us ! That would be a good thing ! david > Mick aka 'sparrowhawk' > > _______________________________________________ > Talk-au mailing list > Talk-au@openstreetmap.org > https://lists.openstreetmap.org/listinfo/talk-au _______________________________________________ Talk-au mailing list Talk-au@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-au