Camera security network design and realization based on PCA facial recognition algorithms

Abstract

Single-Camera Video Surveillance (SCVS) has mainly been applied in the passive monitoring. Video surveillance covers small area. The video information is not sufficient without the joint tracking capability. Currently poor performance in most similar networks is mainly due to unintelligence and low real-time. In this paper we proposed a distributed smart multi-eyes camera network by utilizing smart phones, which presents not only the schemes but also its concrete realization. Using the smart camera network with good self-organization in time, the target would be captured and tracked in good manner and the running track for the target can be restored and recovered anytime. This work is aimed at utilizing unused mobile phones to form a security network. The contribution is aimed at improving the accuracy stranger invasion detection. In order to finish the task of recognizing faces of distant targets, we use PCA-based facial recognition algorithm to enhance the effective resolution. We list different approaches of increasing the resolution and use the best approach to handle different situation. In the mean time we designed and realized the overall structure of the security network including database module design, client terminal design and monitoring node design, forming a robust and energy saving network. This paper can provide valuable reference for the design of the practical security network which is currently in high social demand.