[OSM-talk] GPS and statistics

[Joerg Ostertag (OSM Munich/Germany)]

On Sunday 05 November 2006 11:00, Jan-Benedict Glaw wrote:
> Hi!
>
> Did anybody ever try to statistically calculate the number of parallel
> tracks within one street (eg. two or three for highways, ...) from GPS
> data?  Maybe I'd first look at the NMEA output to see if there are
> enough decimals in the GPS receiver's output at all...

This depends on your precission you currently have. Which depends on the 
position and reception of satelites. But normally if you have a bunch of 
tracks for the some road driven. You can visually(not automatically) see how 
many lanes there are. Problem is that even with a precission of 7 meters the 
tracks for the different lanes do overlap and make it difficult for an 
automatism to find out how many there are. One more point you should keep in 
mind. Som Highways(Autobahn, ..) have a small road/Footpath directly nearby 
(sometimes only 5 meters away). These have to be found too.

just my quick thought on this.

-
Joerg