A comparative analysis to the array database technology and its use in flexible VCI derivation

Recently scientific data are increasingly managed by array database technologies such as SciDB and Rasdaman. Some work has shown that array database technologies are particularly useful to support multi-dimensional data management and analysis. In the geospatial domain, remote sensing images, as a kind of multi-dimensional scientific data, could certainly take advantage of array database technologies. This paper investigates existing array database technologies and evaluates the pros and cons of several typical solutions for managing remote sensing images. It compares the running environments, data structures, APIs, and query languages of three solutions: SciDB, Rasdaman, and Oracle GeoRaster. A benchmark experiment is also designed to test the performance of these three solutions. The results show that Array Database technologies provide more flexibility in managing and analyzing remote sensing images than the traditional relational database solutions. Finally, an study case on VCI derivation for crop vegetation condition monitoring is presented based on Rasdaman database to demonstrate the powerful and flexible capability of in-situ image data processing.