Swissgrild CH1903 is off by up to 3 meters (does not use data set of local corrections )

[bastiK]

To calculate CH1903, 2 steps are required:

(1) WSG84 -> CH1903+ (using a formula)

(2) CH1903+ -> CH1903 (using a grid / triangulation file)

Step (1) if fully implemented. However step (2) is missing. It introduces corrections of up to 3 meters (plus false easting/northing of 2000000/1000000) and accounts for the historical deviations from the modern system.

To fix this, we could:

• Use the "official" triangulation file and FINELTRA algorithm.
• Use the NTv2 grid file provided by swisstopo. This is an approximation, but should be in general good enough for osm use.

Link:  http://www.swisstopo.admin.ch/internet/swisstopo/en/home/topics/survey/lv03-lv95/chenyx06.html

[bastiK]

see also #5327

[bastiK]

triagulation data (binary format)

[bastiK]

On the swisstopo website only NTv2 grid is available, but the link to the triangulation file is dead.

I sent an email to swisstopo and received the file in binary format. (The link is still not working.)

[bastiK]

@sbrunner: Do you like to pursue this? I haven't found any documentation on the low-level format of the binary file.

[sbrunner]

Hi,

Yes, I plan to see what is possible to do but not right now, (actually I'm working on better support of OSM file format in OpenLayers ;-) )

CU, Stéphane

[bastiK]

Ok, if you are able to decipher the binary format, let me know!

Sebastian