Realizing Video-Surveillance on Wireless Mesh Networks: Implementation Issues and Performance Evaluation

Abstract

Video-surveillance systems are currently undergoing a transition from traditional analog solutions to digital ones. This paradigm shift has been triggered by technological advances as well as increased awareness of the need for heightened security in particular vertical markets such as government and transportation. Compared with traditional analog videosurveillance systems, digital video-surveillance offers much greater flexibility in video content processing and transmission. At the same time, it can also easily implement advanced features such as motion detection, facial recognition and object tracking. Many commercial companies now offer IP-based surveillance solutions. This chapter starts from an experience of deployment of a prototype of a large-scale distributed video-surveillance system that the authors’ research group has realized as a common testbed for many research projects. It consists of sixty video cameras distributed over the campus of the University of Catania, transmitting live video streams to a central location for processing and monitoring. Deployment and maintenance of large-scale distributed video-surveillance systems is often very expensive, mainly due to the installation and maintenance of physical wires. The solution chosen in order to significantly reduce the overall system costs, while increasing deployability, scalability, and performance is the use of wireless interconnections (Collins et al., 2001); (Chiasserini & Magli, 2002). With this in mind, the basis idea is to apply multi-hop wireless mesh networks (WMN) (Akyildiz et al., 2005); (Karrer et al., 2003); (Bhagwaty et al., 2003); (Tropos) as the interconnection backbone of a wireless video-surveillance network (WVSN). The proposed architecture is shown in Fig. 1. As we will see in Section 2, it fits in well with the structure of a WMN (Draves et al., 2004), where traffic sources are networked digital video cameras, while the nodes of the WVSN are fixed and wirelessly interconnected to provide video sources with connections towards a video proxy with processing and filtering capabilities. Video proxies are typically located in the wired network. Nevertheless, implementing an intelligent, scalable and massively distributed videosurveillance system over wireless networks remains a research problem and leads to at least the following important issues, some of which have been raised in (Feng et al., 2001):