BUAS: A Knowledge-based Representation of the Border Surveillance Domain

Ioannis Kontopodis, Francisco Javier Iriarte, Peter Leskovský, Jorge García, Oihana Otaegui

Abstract: In this study, we propose a methodology for developing a knowledge-based model that defines the entities of the land border surveillance domain and establishes potential associations among them. A practical ontology was devised to be integrated into a video analysis system, operating on data generated by visual sensors. Its deployment aims to infer complex scenarios and determine advanced spatio-temporal relationships between events. In that way, hidden, upper-level knowledge will be exposed as a set of low-level events ordered in a semantic sequence based on their causal relations. We analyze the core concepts and the general framework of the proposed model, and we spell out the data flow within the ontology. Some crucial parts of the introduced approach are based on the EUCISE-OWL ontology architecture, which was created for a similar purpose in the maritime domain. Additionally, we present the methodology of incorporating our model into a real-time data pipeline with the intention of raising an alert upon the occurrence of an agglomeration of persons. We break down the main components of the system that lead to the population of the ontology, and we semantically decompose the reasoning rules aggregated within the relational database. Finally, the developed ontology is evaluated in terms of expressiveness and completeness, and its major metrics are compared with the EUCISE-OWL ontology,