一种基于无线视觉传感器网络的分布式多目标跟踪滤波算法  

A Distributed Multi-Target Tracking Filter Algorithm Based on a Wireless Visual Sensor Network

作　　者：涂蓝 

机构地区：[1]江西理工大学计算机科学与技术系,江西赣州341000

出　　处：《科技广场》2014年第10期18-24,共7页Science Mosaic

摘　　要：本文提出基于积分卡尔曼滤波的无线视觉传感器网络的分布式目标跟踪算法,用于从多个监控数码相机中汇总跟踪结果,然后通过中央处理器处理数据,返回到各个数码相机中,从而优化目标跟踪。无线视觉传感器网络是一个重叠视觉领域里的一个多监控摄相机网络,重叠视觉领域里有同样的目标经过,通过交换它们当地的目标位置和速率的估计值,在估计的过程中,每个加入的摄相机通过一致性算法汇总收到的状态。仿真结果表明,多个数码相机对目标预测的真实状态值要比单个数码相机预测值更精确。The task presents a distributed wireless sensor network based on cubature Kalman filter of visual target tracking algorithm for tracking results from multiple surveillance cameras and processing the data by central processor and returning them to the digital camera in order to optimize target tracking. Wireless network is a mul-ti-vision monitor camera network in an overlay visual sensor where there passes the same goal by exchanging their local target position and the estimation of velocity. During the procedure, each camera in the network gather re-ceived status through consensus algorithm. Simulation results show that multiple digital cameras have a more pre-cise result for the prediction of the true state of the target than a single one.

关 键 词：无线视觉传感器网络 积分卡尔曼滤波 一致性算法 多目标跟踪 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]