Evaluation of a NoSQL Database for Storing Big Geospatial Raster Data
Nicole Hein, Jörg Blankenbach · GI_Forum · 2021
Database systems capable of efficiently storing geospatial data are widespread. However, recent developments in earth observation systems, remote sensing, mobile mapping, and crowd sourcing lead to large amounts of geospatial mass data that can hardly be handled efficiently with the existing solutions. Especially large geospatial raster data require novel concepts for well-organized data storage. A concept for storage of large geospatially and temporally referenced image data using the NoSQL graph database system Neo4j as a research subject of the project “RiverView®” is introduced. New strategies and access structures have been developed to ensure the persistence and performant access to image data in Neo4j. These strategies are compared with the up-and download performance of the widespread Rasdaman array database system.