Hi all,
You might be interested to know about a system we have been working on
for storing large amounts of spatio-temporal data. The system focuses
on ease of data access: after ingesting a dataset, it is available as
one large image from which you can extract subsections. This is done
using GDAL for import and export, and NetCDF-Java for viewing and
processing. The internal storage format is tiled NetCDF 4 classic.
The system was called the Raster Storage Archive during development;
now we call it SpatialCube. It's free software with commercial support
available [1]. Here are some videos explaining what it does and the
motivation behind it:
Short video (2:34)
http://www.youtube.com/watch?v=JuDoyVLuU44
Long video (13:07) - more detail; includes examples of generated NCML
http://www.youtube.com/watch?v=YZwJW1fABnk
Perhaps of particular interest to this list: SpatialCube includes an
image processing system built on top of NetCDF-Java called rsaquery.
The videos introduce the query engine; details of the design can be
found in the repository on GitHub [2]. The design of the storage
system is also documented there [3].
Questions, comments and collaboration are of course welcome.
Cheers,
Alex
--
Alex Fraser
Software Engineer
VPAC Innovations Pty Ltd.
[1] http://vpac-innovations.com.au/work/spatialcube
[2] https://github.com/VPAC/rsa/blob/master/doc/design-rsaquery.md
[3] https://github.com/VPAC/rsa/blob/master/doc/design.md
