The BigDataCube project has the goal of advancing the paradigm of the “data cube” – in other words, ready-to-analyse spatiotemporal raster data – from the status of a science prototype to a pre-competitive earth-observation service.
Data cubes are three- or more dimensional data structures that allow rapid and direct analysis of large amounts of data. One application area is earth observation data. The growing number of remote sensing satellites, such as the ESA Sentinel missions, increases the amount of available data by many terabytes every day. For many analyses, however, only a limited region may be needed, or only scenes collected under particular recording conditions. Instead of downloading and completely analysing many large-scale images in order to obtain that selection, various filters are specified in a procedure similar to a database query and the results are made available to the user in the form of a data cube. In this way analyses and calculations can be made only for the required time period, region and parameters.
As part of the project, DLR’s Maritime Safety and Security Lab in Bremen will adapt for use in data cubes its existing automated processors for extracting wind and wave information from SAR (Synthetic Aperture Radar) images. This makes it possible to easily tap available earth observation data to answer questions like “What was the peak wave height on the German North Sea coast in 2017?” or “What is the average wind velocity at the planned location for an offshore wind park?”
The “rasdaman” data cube technology developed in Germany is to be installed on CODE-DE as well as in the commercial cloud environment of CloudEO as part of the project. This will make it possible to offer examples of analytic services and establish relevant links. The project is supported by the Federal Ministry for Economic Affairs and Energy (BMWi) and will run for 18 months. It is being coordinated by Jacobs University (Bremen); other partners are rasdaman GmbH (Bremen) and CloudEO AG (Munich).