Ontology-Based Intelligent Security Framework for Smart Video Surveillance

Abstract

Modern smart surveillance systems have to deal with heterogeneous surveillance devices and the huge size of the video surveillance data. The surveillance data also contains sensitive and personal information and, hence, the security of the surveillance data is a demanding requirement. Owing to the large data size of the surveillance videos, conventional security measures cannot possibly be applied to all of the video surveillance data and, consequently, this presents a significant challenge within the resource constraints and dynamics of the emerging Internet of Things (IoT). Therefore, appropriate and automated intelligent systems need to be adopted for management, secure transmission, secure storage, and efficient retrieval of the ever-increasing surveillance video data. In this paper, an intelligent surveillance security framework for videos is proposed. The proposed framework for real-time video processing is integrated with surveillance video semantic concepts to provide an effective and efficient solution for selection, implementation, and manipulation of heterogeneous surveillance devices. The aim is to create a core, Secure Smart Surveillance Ontology (SSSO) to represent and share the common vocabularies for secure communication and storage of surveillance video for different surveillance devices, with an ontology being the explicit formal representation of basic categories within the domain. Firstly, different security levels are defined with respect to heterogeneous devices on the basis of their respective storage capacity to ensure sufficient encryption details with minimal resources used. Secondly, in the proposed ontology, device-specific security concepts are linked with the surveillance video concepts. After that the proposed SSSO is integrated within the security framework. Overall, the proposed security framework allows relatively fast indexing and retrieval along with device-specific security.