Big EO data / Big Earth data requires new and innovative methods for access, storage, analysis, understanding and sharing of data. While in other big data domains the volume, velocity and variety is a challenge on its own, remote sensing data require intelligent techniques for transforming pixel numbers into valuable information. Only when the computer-based transformation from data into information succeeds, the data can be a source for providing reliable, multi-temporal and globally available geospatial information.
The research activities in the OBIA working group aim at automatic semantic image understanding, methods for combined spatial and temporal analysis of satellite image time series directly in large databases, as well as analysis of uncertainty and inhomogeneity in global data coverage and data quality.
Example developments, which are spanning over multiple projects, are the ImageQuerying (IQ) system as a prototype of an EO-image understanding and semantic querying system, and the Sentinel-2 Dashboard, as a tool to explore global Sentinel-2 Metadata interactively.