基于视频传感器网络的高效分布式跟踪方法  

High efficient distributed tracking method based on video sensor networks

作　　者：方武[1,2] 王典洪[1] 王勇[1] 

机构地区：[1]中国地质大学(武汉)机电学院,湖北武汉430074 [2]苏州经贸学院信息系,江苏苏州215009

出　　处：《传感器与微系统》2014年第7期17-19,23,共4页Transducer and Microsystem Technologies

基　　金：国家自然科学基金青年基金资助项目(41202232);江苏省高级访问工程师项目;武汉市青年科技晨光计划资助项目(2014070404010209)

摘　　要：近年来为减少无线视频传感器网络的通信数据量从而节约能耗,一些研究引入高效分布式算法对目标进行跟踪。然而,常用的分布式跟踪算法采用信息矩阵进行数据融合来提高跟踪精度,数据通信量虽然比传递图像小,但依然比较耗能。针对这种情况,提出一种基于残差编码的一致性卡尔曼滤波跟踪方法,对信息矩阵采用符号编码策略,从而大量减少视频传感器节点之间数据交换量,实验证明:该方法在保证跟踪效果的同时可减少70%左右的能量消耗。In order to reduce energy consumption of nodes by reducing amount of data for communication of wireless video sensor networks（ WVSNs） in recent years,high efficient distributed target tracking methods are proposed.However,distributed consensus tracking method is commonly used,which uses information matrix for data fusion to improve tracking precision,although communication amount is lower than transmitting image,it is still energy consuming.Aiming at this situation,present a Kalman consensus filtering（ KCF） tracking method based on residual coding（ RC）,sign coding strategy is used to information matrix which can dramatically reduce datas transmission amount between video nodes,experiments show that the method can effectively reduce energy consumption about 70 % and assure tracking effect at the same time.

关 键 词：残差编码 分布式跟踪 无线视频传感器网络 数据融合 

分 类 号：TP317.4[自动化与计算机技术—计算机软件与理论]