Adaptive home surveillance system using HTTP streaming

In this paper, we investigate the use of HTTP streaming in home surveillance system. First, we propose a dynamic resource allocation framework among the cameras. Under a set of constraints, e.g. link bandwidth and delay, our resource allocation solution decides which video versions will be requested from which cameras in order to maximize the weighted sum of utility for all cameras. The optimization, which is performed at the client, is based on a utility model for each video stream that takes into account not only perceptual quality but also the delay of a video stream. Video content is obtained through HTTP request/response transaction between the client and the cameras. The experiment results show that the proposed method is able to effectively improve the overall utility while still supporting low end-to-end delay and good perceptual quality for important cameras.

I. Introduction

Video-based surveillance system [1], [2] has been widely applied in many places such as airports, train stations, banks, etc. The general purpose of a surveillance system is to help detecting and preventing illegal and unwanted activities. Recently, due to the widespread increase of security threats, e.g. terrorism, the use of surveillance technologies has become an essential requirement for organizations and individuals. However, most existing surveillance systems still depend on human operations. That raises the issue of having a surveillance system which effectively supports human operators to make proper decisions.