[geotk] Transform from EPSG 4979 to 3D CRS

[Martin Desruisseaux]

Hello

Le 04/05/11 19:56, Aaron Braeckel a écrit :
> Are there any authorities (ex: EPSG) that define a general-purpose
> ellipsoidal vertical CRS analogous to EPSG:5714?

I'm not aware of any code in the EPSG, CRS, AUTO or AUTO2 namespace. Actually 
the ISO 19111 specification explicitely forbives "ellipsoidal height" as a 
standalone vertical CRS, for conceptual reason. They said that an ellipsoidal 
height makes sense only when associated to a geographic CRS (at the difference 
of other kind of height which can be used alone). Geotoolkit.org can handles 
standalone ellipsoidal heights for pragmatic reason, but this is a departure 
from ISO 19111 specification.

ISO 19111 handles ellipsoidal heights in a very special way. We can get an 
ellipsoidal height only as a component of a 3D Geographic CRS. There is a EPSG 
code for a 3D Geographic CRS (WGS84) with vertical ellipsoidal height: EPSG:4327.

For other kind of CRS (projected...), an alternative may be to use WKT. The WKT 
specification has been created before ISO 19111, at a time where standalone 
ellipsoidal height were not considered "bad". So it is possible to express them:

    VERT_CS["Ellipsoidal height",
       VERT_DATUM["Ellipsoidal height", 2002],
       UNIT["metre", 1.0],
       AXIS["Ellipsoidal height", UP]]


Except for the above WKT, if we want to be strictly ISO/OGC compliant, the main 
alternatives I can see are:

    *

      Help OGC to restart the referencing SWG, and propose to allow standalone
      ellipsoidal height in some cases.

    *

      Finish the "ellipsoidal to geoidal" transformation in Geotk. But even if
      this functionality was implemented, it would have significant cost at
      runtime, which may not be always desired.


Martin

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osgeo.org/pipermail/geotoolkit/attachments/20110505/9c04832e/attachment.html