Architecture and Scheduling Method of Cloud Video Surveillance System Based on IoT

摘要

To realize conveniently deployed video surveillance applications, this paper designs a cloud service system employing ubiquitously available IoT nodes. Considering limited capacity of each IoT node, this paper first describes the system architecture and operation procedure for application requests, and introduces the design of scheduler's function and typical video processing algorithms. Further, for decreasing transmission conflicts among video/image processor nodes, this paper proposes a scheduling methods based on Genetic Algorithm to rationally utilize the cooperative IoT nodes. Simulation results show that, compared with common methods such as random scheduling and opportunity-balanced scheduling, this method yields much smaller processing delay and transmission delay, together with higher packet delivery ratio.