Scalable Spatio-temporal Analysis on Distributed Camera Networks

Abstract

Technological advances and the low cost of sensors enable the deployment of large-scale camera networks in airports and metropolises. A wellknown technique, called spatio-temporal analysis, enables detecting anomalies such as an individual entering into a restricted area without permission. Spatio-temporal analysis requires a large amount of system resources to infer locations of occupants in real-time. In particular, state update becomes a bottleneck due to computation and communication overhead to update possibly large application state. In this paper we propose a system design and mechanisms for scalable spatio-temporal analysis. We present a distributed system architecture including smart cameras and distributed worker nodes in the cloud to enable real-time spatio-temporal analysis on large-scale camera networks. Furthermore we propose and implement a couple of selective update mechanisms to further improve scalability of our system by reducing the communication cost for state update.