A Flexible and Scalable Cooperative Surveillance Technique Using Integrated Ontology Reasoning

Abstract

In large-scale smart camera networks, cooperation among devices is required for continuous tracking of targets and higher level reasoning. A large amount of multimedia data with derived metadata is generated and transferred among devices. In this paper, we design a large-scale surveillance system which consists of smart cameras. It complies with the standard specification to ensure interoperability among cameras and flexibility regarding integration of new devices and services. Surveillance data contained in them is integrated and structured according to the ontology, and useful context information can be derived. This paper introduces how to build surveillance knowledge base, import relevant data from other devices, and annotate data on interoperable framework which accommodates to the standard. The annotation process provides an impetus to the improvement of knowledge over time. We define a representative reasoning architecture that provides location-based context induction, and implemented in our test bed site to show superiority in large-scale surveillance.