[gdal-dev] Interpolation used for egm96_15.gtx when using RPC_DEM_APPLY_VDATUM_SHIFT

[Yann-Sebastien Tremblay-Johnston]

Hi Even,

To start with, you need the the dataset pointed by RPC_DEM is a compoundCRS
> with a vertical CRS properly defined


Thanks, I initially didn't understand why this wasn't working as the SRTM
DEM I am using does not have a compoundCRS. I worked around this by
creating a VRT with the correct compoundCRS.

 Interpolation of the geoid grid is done by PROJ itself, and this is
> bilinear interpolation


Awesome, good to know. Is there anyway of configuring this?

Thanks for the quick response.

Sebastien

On Fri, Apr 17, 2020 at 10:09 AM Even Rouault <even.rouault at spatialys.com>
wrote:

> Yann-Sebastien,
>
>
>
> > I spent a bit of time learning about vertical datum shifts when
> performing
>
> > RPC coordinate transformations using a DEM this week, and I was curious
>
> > about how the egm96_15.gtx file is interpolated. I was able to find in
>
> > gdal_rpc.cpp that the required vdatum shift is calculated using an OGR
>
> > Coordinate Transformation between the DEM spatial reference and a wgs84
>
> > ellipsoid spatial reference, but I’m not familiar enough with OGR/PROJ to
>
> > understand exactly how the gtx file is sampled.
>
>
>
> To start with, you need the the dataset pointed by RPC_DEM is a
> compoundCRS with a vertical CRS properly defined and an existing
> transformation between that vertical CRS and WGS84, so if it is something
> like EPSG:XXXX+5773, that will be fine.
>
> Interpolation of the geoid grid is done by PROJ itself, and this is
> bilinear interpolation.
>
>
>
> Even
>
>
>
> --
>
> Spatialys - Geospatial professional services
>
> http://www.spatialys.com
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/gdal-dev/attachments/20200417/f4b05034/attachment.html>