Re: [Gpsbabel-misc] Accuracy of BNG > GPX (WGS84) transformations

[tsteven4 <tst...@gm...>]

I suspect but have not proven this is because we use a abridged 
molodensky transformation using the 3 parameters dX, dY dZ values shown 
below.  The citation claims this is only accurate to +/- 20m.  We don't 
use PROJ.4, and we don't have more complex transformations available.

See Transformations in GDAL/OGR (edina.ac.uk) 
<https://digimap.edina.ac.uk/help/gis/transformations/transform_gdal_ogr/> 
which says

> PROJ.4 contains a large number of transformations and parameters. 
> However by default OGR uses a 3 parameter shift because it covers the 
> largest available area -see blogpost on how transformations are chosen 
> in Proj4. 
> <http://fwarmerdam.blogspot.com/2010/03/in-last-few-weeks-i-believe-i-have-made.html>However, 
> this is only accurate to +/- 20 metres over the whole of the UK. It is 
> equivalent to the ArcGIS transformation called OSGB_1936_To_GS_1984_1 
> and has parameters of dX: 375, dY-111, dZ 431

On 8/5/2024 11:37 AM, Nims via Gpsbabel-misc wrote:
> Hi Robert,
> I've been converting Universal CSV lists of BNG coordinates to GPX XML 
> files with GPSBabel and am wondering about the accuracy of the 
> conversion. Here's my test input which lists a couple of trig points 
> (triangulation stations) in the county of Dorset:
> name,comment,bng,date
> Badbury Rings Trig Point,Test Point 1,ST 96552 03097,2024/08/05
> Spettisbury Rings Trig Point,Test Point 2,ST 91437 01894,2024/08/05
> The trig points are marked on British Ordnance Survey maps and also 
> visible in satellite images, allowing me to check accuracy in various 
> mapping apps.
> Using the GPSBabel GUI (v.1.9.0) under Windows 10, the file is 
> processed successfully with the following feedback:
> gpsbabel -w -i unicsv -f G:/Trig Points.csv -o gpx -F G:/Trig Points.gpx
> Translation successful
> This is the output:
> <?xml version="1.0" encoding="UTF-8"?>
> <gpx version="1.0" creator="GPSBabel - https://www.gpsbabel.org" 
> xmlns="http://www.topografix.com/GPX/1/0">
>   <time>2024-08-05T16:54:46.903Z</time>
>   <bounds minlat="50.816463984" minlon="-2.122941465" 
> maxlat="50.827336109" maxlon="-2.050345019"/>
>   <wpt lat="50.827336109" lon="-2.050345019">
>     <time>2024-08-04T23:00:00Z</time>
>     <name>Badbury Rings Trig Point</name>
>     <cmt>Test Point 1</cmt>
>     <desc>Test Point 1</desc>
>   </wpt>
>   <wpt lat="50.816463984" lon="-2.122941465">
>     <time>2024-08-04T23:00:00Z</time>
>     <name>Spettisbury Rings Trig Point</name>
>     <cmt>Test Point 2</cmt>
>     <desc>Test Point 2</desc>
>   </wpt>
> </gpx>
> All well and good but when I check the generated coordinates in 
> mapping apps, they are a few metres off target. Furthermore, they 
> differ from coordinates generated by various online BNG > WGS84 
> conversion tools which are themselves consistent. For example:
> British Geological Society
> https://webapps.bgs.ac.uk/data/webservices/convertForm.cfm
> Badbury Rings Trig Point = 50.827400 , -2.050323
> Spettisbury Rings Trig Point = 50.816527 , -2.122918
> Grid Reference Finder
> https://gridreferencefinder.com
> Badbury Rings Trig Point = 50.827400 , -2.0503234
> Spettisbury Rings Trig Point = 50.816527 , -2.1229184
> Mapserve
> https://www.mapserve.co.uk/conversion-tools
> Badbury Rings Trig Point = 50.8273995082622 , -2.05032333478781
> Spettisbury Rings Trig Point = 50.8165274566942 , -2.1229183177272
> The difference isn't huge but it seems to be more than a rounding 
> error. So, I'm wondering what might need to be done to harmonise the 
> GPSBabel conversion with these others.
> Many thanks,
> Nims
>
>
> _______________________________________________
> Gpsbabel-misc mailing listhttp://www.gpsbabel.org
> Gps...@li...
> To unsubscribe, change list options, or see archives, visit:
> https://lists.sourceforge.net/lists/listinfo/gpsbabel-misc