双目视觉的弱点动目标粒子滤波跟踪定位研究  

Study on dim moving point target’s particle filter tracking and orientation based on binocular vision

作　　者：周同驰[1] 艾斯卡尔.艾木都拉[1] 杨强[1] 王荣栓 

机构地区：[1]新疆大学信息科学与工程学院,乌鲁木齐830046 [2]天津市雅浪通讯,天津300384

出　　处：《计算机工程与应用》2012年第9期185-188,共4页Computer Engineering and Applications

基　　金：国家自然科学基金(No.60507005);新疆维吾尔自治区高校青年教师科研启动基金项目(No.XJEDU2008S14)

摘　　要：在研究红外图像序列的弱点动目标粒子滤波跟踪算法和双目立体视觉摄像机标定算法的基础上,基于双目视觉设计了双目图像序列弱点动目标的跟踪、空间定位系统。仿真实验表明:对粒子基于估计参数后的密度函数分配权重的算法,提高了图像跟踪精度,采用线性三角定位方法,有效地实现目标空间定位。This paper designs an experimental system for tracking and spatial localization of a dim moving point target in binocular im- age sequences based on the research of particle filter tracking algorithm and the binocular vision camera calibration algorithm. Experi- mental results show： after estimating the density function＇ s parameters, the algorithm which distributes the weight to particle according to PDF advances the precision of image tracking; the method which adopts linear triangulation principle to extract the target＇ s spatial lo- calization is effective.

关 键 词：红外图像序列 弱点动目标 粒子滤波 双目视觉 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]