OntoRaster: Extending VKGs with Raster Data

Proc. of the 8th Int. Joint Conf. on Rules and Reasoning (RuleML+RR 2024). Volume 15183 of Lecture Notes in Computer Science. 2024. Best student paper award (ex-aequo).

The Virtual Knowledge Graph (VKG) paradigm facilitates access to large heterogeneous data sources by leveraging an OWL 2 QL ontology representing the domain knowledge and a set of declarative R2RML mapping assertions. We are interested in heterogeneous data sources consisting of relational data together with spatial geometrical data (a.k.a. vector data) and large multidimensional raster data. The latter forms of data pose a significant challenge for traditional DBMSs to manage effectively and are instead efficiently processed by tailored array database management systems (ArrayDBMSs). To query such data within the VKG paradigm, we propose a novel framework, called OntoRaster, that allows for integrated query processing of relational, raster, and vector data, by keeping each type of data in the system tailored for their efficient processing, while minimising costly data-transfer operations. In OntoRaster, we devised custom raster functions extending SPARQL to query raster data efficiently and developed mechanisms for delegating their computation to the ArrayDBMS. We have implemented the whole framework as an extension of the state-of-the-art VKG system Ontop and have demonstrated its effectiveness and efficiency through a curated case study.

@inproceedings{RuleML-RR-2024-ontoraster,
   title = "OntoRaster:  Extending VKGs with Raster Data",
   year = "2024",
   author = "Arka Ghosh and Albulen Pano and Guohui Xiao and Diego
Calvanese",
   booktitle = "Proc. of the 8th Int. Joint Conf. on Rules and Reasoning
(RuleML+RR 2024)",
   pages = "108--123",
   volume = "15183",
   publisher = "Springer",
   series = "Lecture Notes in Computer Science",
   doi = "10.1007/978-3-031-72407-7_9",
   note = "Best student paper award (ex-aequo)",
}