多源测量信息融合的弱点动目标跟踪算法  

Dim moving point target tracking algorithm based on multi-source measurements information fusion

作　　者：周同驰[1] 艾斯卡尔.艾木都拉[1] 王荣栓 黄冬梅[3] 

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 [3]乌鲁木齐铁路局电务段,乌鲁木齐830011

出　　处：《计算机工程与应用》2012年第2期182-184,226,共4页Computer Engineering and Applications

基　　金：国家自然科学基金(No.61065001);新疆维吾尔自治区科技厅少数民族特殊培养计划项目(No.201023116)

摘　　要：在复杂背景下,为提高微弱点动目标跟踪系统的抗干扰能力,从不同信噪比的多红外图像序列出发,基于新的融合跟踪结构(增加局部处理器),提出了一种模糊逻辑的多源信息测量融合跟踪算法:各局部处理器对原始测量进行恒虚警和自适应检测,将判决后的测量送入融合中心,融合中心对测量作模糊逻辑判决融合,转化为虚拟单序列测量,采用PDA卡尔曼滤波算法跟踪。实验结果分析表明,该融合跟踪算法与单序列相比,具有较高的跟踪精度、稳定性,避免了单图像序列跟丢。In order to improve anti-jamming of a dim moving point target tracking system under complex backgrounds,based on the new fusion tracking frame（add the local processor）,tracking algorithm which is multi-source measurements information fuzzy logic fusion is proposed in different signal-to-noise conditions：local processor judges the original measurements by CFAR and adaptive detection threshold,transmits the measurements to fusion center;measurements which are fused by the fuzzy logic method are transformed into single virtual image sequence＇s measurements;adopts PDA-Kalman algorithm to track target.Compared to the single sequence,experimental analysis and results show that the fusion tracking algorithm has high accuracy and good stability,avoids losing the single sequence track.

关 键 词：多红外图像序列 弱小目标 测量融合 概率数据关联滤波 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]