#5032 closed enhancement (fixed)
HDF5 driver fails to read ODIM georeferencing method
Reported by: | sjurk | Owned by: | warmerdam |
---|---|---|---|
Priority: | normal | Milestone: | 1.11.0 |
Component: | GDAL_Raster | Version: | 1.9.2 |
Severity: | normal | Keywords: | HDF5 proj |
Cc: | antonio |
Description
Using Windows 7 and precompiled libraries downloaded from http://www.gisinternals.com/sdk/, GDALINFO in GDAL 1.7.3 reports some ellipsoid/datum info, whereas GDALINFO in GDAL 1.9.2 does not. Does 1.7.3 find the ellipsoid/datum info in the file, or is WGS84/WGS84 assumed by default in 1.7.3, and not in 1.9.2?
The file is a HDF5 ODIM file containing weather radar data, exported from the Norwegian met office (www.met.no). ODIM is the Opera Data and Information Model (http://www.eumetnet.eu/opera).
The GDALINFO output also differs in that 1.9.2 precedes some of the ODIM metadata tags with dataset1_, and reports the Conventions=ODIM_H5/V2_1 tag. 1.7.3 does not.
Neither version finds the projection info. The string: where:projdef=+proj=utm +zone=32 +ellps=WGS84 +datum=WGS84
or
where:projdef=+proj=aeqd +lat_0=63.536783 +lon_0=10.204000 +R=6371000 +datum=WGS84
appears in two different versions of the file (both enclosed).
Is it possible to look for strings starting with proj= in all metadata items, and extract the info if meaningful?
Attachments (2)
Change History (5)
by , 11 years ago
Attachment: | norsa.ss.ppi-00.5-dbz.aeqd-1000.20070601T000039Z.hdf added |
---|
by , 11 years ago
Attachment: | norsa.ss.ppi-00.5-dbz.utm32-1000.20070701T000039Z.hdf added |
---|
Radar file in UTM zone 32 projection
comment:1 by , 11 years ago
Cc: | added |
---|
comment:2 by , 11 years ago
Component: | default → GDAL_Raster |
---|---|
Milestone: | → 2.0 |
Resolution: | → fixed |
Status: | new → closed |
Summary: | HDF5 driver fails to read +proj= string, 1.7.3 and 1.9.2 behave differently → HDF5 driver fails to read ODIM georeferencing method |
trunk r25972 "HDF5: add support for ODIM H5 georeferencing method (#5032)"
There was never a regression. The WGS84 SRS reported in GDAL 1.7.3 was wrong, and later versions just remove that wrong information. The above commit should implement a proper decoding of the georeferencing info, although some tests should be done to check it is OK. See comment in the code for the uncertainty in the decoding.
comment:3 by , 10 years ago
Milestone: | 2.0 → 1.11.0 |
---|
Radar file in Azimuthal Equidistant projection