Re: [Gpsbabel-misc] Source code to imporove distance calculation

[Ron P. <ro...@pa...>]

David W. Bray wrote:

>
> I found that the OpenOffice .csv file gave a distance calculation.
> However is not accurate.  I have fought this problem several times so 
> I would like to give you my GreatCircle routine which appears to be 
> accurate for small distances.
>
> I extracted the lat/lon via GPSBabel to a text file. I then used my 
> routine to add up the individual distances. The results were:
>
> Delome route: 224.5 miles
> Converted text file: 227.1 miles
> GPSBabel.csv: 365.9 miles
>
I think you'll find that the value GPSBabel gives is in kilometers.  It 
can generate path distance in miles, but the OpenOffice stylesheet 
doesn't use that value.  Performing the appropriate conversion yields a 
value that is quite close to the one you computed.

The formula we use to compute great-circle distance is 
acos(sin(lat1)*sin(lat2)+cos(lat1)*cos(lat2)*cos(lon1-lon2)) which gives 
an angle in radians.  We then multiply that by the equatorial radius of 
the earth as specified in the WGS84 datum (6378137 meters) which isn't 
quite accurate, but not much less accurate than pretending the earth is 
a sphere to begin with.