Problem with COSMO data georeference

[mguerecheau]

Hi, We saw that when we open COSMO data (1B-DGM level, Spotlight sensor mode) that are delivered in tiff, WGS 84, the GDAL software does not read the header file correctly as the georeference system WGS 84 is not detected.

DO you know why and would it be possible to correct it ?

Thank you very much,

Marine

[Even Rouault]

Probably because there are no geotiff tags in those files ? Could you provide a link to such files ?

[mguerecheau]

header CSK Spotlight data 1B-DGM

[mguerecheau]

Thank you very much for your fast answer. Please find attached the header file, indicating geotiff format.

Do you know from where does the problem come ?

Thank you very much,

Marine

[Even Rouault]

GDAL does not read such .h5.xml files. It will only try to read the information from the geotiff header which is inside the tiff file itself. You could try the "listgeo" utility provided with the geotiff library on the .tif file, and paste the output of it. What is puzzling is also the .h5.xml extension that would seem to indicate this is a HDF5 file...

[pmallas]

Cosmo SkyMed's default output is HDF5. They implemented geotiffs at CSTARS's request several years ago. Of course, the geotiff cannot hold all the meta data that the hdf5 can. We suspect they used h5dump to extract the metadata from the HDF5 to create the xml file. However it was created, the xml file contained corrupted/truncated metadata. I recommend using only the hdf5 format for Cosmo if you want reliable metadata.

[Jukka Rahkonen]

Metadata was server in HDF5 format for GeoTIFF image. I consider that the metadata are faulty, not GDAL.