Seeking someone knowledgeable in the ways of crafting GDAL drivers to help with that part of the effort.
The enclosed program extracts (rescues!) raster data from an ArcGIS File Geodatabase into a GeoTIFF file.
List the numbers and names of rasters in the geodatabase using
./arc_raster_rescue.exe <path/to/geodatabase.gdb/>
Extract to a GeoTIFF using
./arc_raster_rescue.exe <path/to/geodatabase.gdb/> <RASTER NUM> <OUTPUT FILE>
The geodatabase path must end with a slash!
On an Ubuntu/Debian system you can obtain these with:
sudo apt install cmake libgdal-dev zlib1g-dev g++
To compile use the standard cmake sequence:
mkdir build
cd build
cmake -DCMAKE_BUILD_TYPE=RelWithDebInfo ..
make
Add -DEXPLORE=ON
to print additional information useful for development to
stderr as the program runs.
An executable called arc_raster_rescue.exe
is produced.
- Turn this into a GDAL driver.
- Improve endian checking.
- Check for other kinds of compressions.
- Improve calculation of raster dimensions.
- Improve geotransform calculations: they're kind of experimental at the moment.
- Switch some identification strings to ENUM values so things run quicker
-
Even Roualt did much of the work figuring out the FileGeodatabase specification. His notes are here and a program he wrote for extractin FGDB data is here.
-
James Ramm did some exploratory work that resulted in determining that the raster data (at least in cases he tested) had been compressed using zlib. (Link)
-
Richard Barnes converted code by Even Roualt into C++ as a base for Arc Raster Rescue and, starting from some notes by James Ramm, continued deciphering the format. He reorganized the code extensively to increase its modularity, determined how to connect the various raster tables together, drafted geotransform and WKT projection extraction capabilities, determined how data types were specified, and produced a working executable to extract raster data into GeoTIFFs. He also wrote this paragraph. :-)
Please cite this software:
Barnes, Richard. 2020. Arc Raster Rescue. Software. doi: 10.5281/zenodo.4128479.