[Proj] +towgs84 approximation error

[Jochem]

Hello Martin,

What you describe about the 80 transformations to NAD27 sound as 80
different frames to me, or one frame that is in need of a NTv2 correction
grid instead of 80 attempts for a 7 parameter transformation.

> if we transform that coordinate to RD, then the location described by the
> new coordinate may be slightly 
> aside the physical object

The RD coordinates of any object in the Netherlands are by definition the
transformed  ETRF coordinates. With a new realisation of ETRS89 the ETRF
coordinates of objects can change. If this change is less than 1 cm, the RD
coordinates change too. If this change is larger than 1 cm we update the
transformation to keep the RD coordinates as close as possible to the 1918
values. 

I would describe this as a series of RD realisations (1918, 2000, 2004 and
2008 frames) which are all connected to ETRS89 realisations by a
transformation. Of course these transformations have been estimated on
stochastic data, however the transformations of the frames since 2000 are by
definition correct (conventional transformations) en will always point to
the same point as the ETRS89 coordinates.

Does that clarify the Dutch situation?

Regards, Jochem 




--
View this message in context: http://osgeo-org.1560.x6.nabble.com/towgs84-approximation-error-tp5313738p5314215.html
Sent from the PROJ.4 mailing list archive at Nabble.com.