[gdal-dev] NetCDF: North is left

[Markus Metz]

Even,

On Tue, Mar 19, 2019 at 4:53 PM Even Rouault <even.rouault at spatialys.com>
wrote:
>
> Markus,
>
> yes this file has unconventionnal indexing of variables, with the fastest
> varying dimension being latitude, which confuses the driver
>
> In https://github.com/OSGeo/gdal/pull/1377, I've made a number of changes
that
> make support of such dataset a bit easier. I didn't go to making the
driver
> autocorrect the swapped axis, but did a number of fixes so that the
> geotransform and Geolocation array metadata are correct (and also improve
the
> geoloc transformer/wrapper to deal with such swapped axis)
>
> With those fixes, the following does the right thing:
> $ gdalwarp -geoloc  NETCDF:"LPRM-
> AMSR2_L3_DS_A_SOILM3_V001_20190301000851.nc4":ts /tmp/out.tif -overwrite

thanks a lot!

But looking at all the datasets with weird georeferencing out there in the
wild, it would be really ambitious to come up with a fix at driver level...
>
> Without it, you can indeed have the expected result by setting GCP

This is what I ended up doing, it works just fine, also in an automated
processing chain.

Markus M

>
> $ gdal_translate \
>    NETCDF:"LPRM-AMSR2_L3_DS_A_SOILM3_V001_20190301000851.nc4":ts \
>    tmp.vrt \
>    -gcp 0 0 -180 90 \
>    -gcp 1800 0 -180 -90 \
>    -gcp 0 3600 180 90 \
>    -gcp 1800 3600 180 -90
>
> $ gdalwarp tmp.vrt out.tif -order 1 --config GDAL_NETCDF_BOTTOMUP 0
-overwrite
>
> 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/20190319/d87ec4b2/attachment-0001.html>