Collaborative Synchronization Mechanism in Wireless Multimedia Sensor Networks

Abstract

The synchronization problem of the wireless transmission in real-time visual surveillance applications was addressed. The distributed architecture of wireless multimedia sensor networks was adopted to obtain the key frames from multi-views synchronously, which has low computational cost and memory requirements. A good strategy for distributed computing was designed to assign tasks to specific smart devices. The main contribution of this architecture was a combination of wireless sensor networks and visual sensors. Image coordinates of detected objects are encapsulated in wireless data transmission. To track 3D localization in compute vision, the collaborative synchronization mechanism was presented to realize the synchronization of the key frames and asynchronization of the packet transmission from multi-views. The received data were divided into groups by the real-time clustering algorithm for 3D localization. Experimental results showed the proposed architecture can implement reliable and efficient indoor monitoring.