基于标签多伯努利滤波器的机动小目标检测前跟踪  

Track-before-detect for maneuvering small targets based on Labeled Multi-Bernoulli filter

作　　者：李淼[1] 龙云利[1] 李骏[1] 安玮[1] 周一宇[1] 

机构地区：[1]国防科技大学电子科学与工程学院

出　　处：《红外与毫米波学报》2016年第5期625-633,共9页Journal of Infrared and Millimeter Waves

基　　金：中国博士后科学基金(2013M532167)~~

摘　　要：标签多伯努利(LMB)滤波器在传统多伯努利滤波器基础上引入标签空间,能够实现真正意义上的多目标轨迹级滤波.文章对红外小目标的运动和量测进行建模,将标签多伯努利应用到红外小目标检测前跟踪领域.在此基础上,为了实现对运动模型时变目标的检测前跟踪,将交互式多模型(IMM)与LMB检测前跟踪算法相结合,提出IMM-LMB检测前跟踪算法.此外,给出了该算法的序贯蒙特卡罗实现.仿真结果表明,所提算法能够从输入的原始图像中直接实现轨迹级多目标检测和跟踪,且能够在线更新多模型概率,更好的适应多机动目标场景.Labeled M ulti-Bernoulli（ LM B） filter was proposed based on classic M ulti-Bernoulli filter by taking label space into account. It is a true trajectory filter in comparison with other methods. The motion model and measurement model of infrared small targets were studied. The LM B Track-Before-Detect algorithm was applied to detect and track infrared small targets. In order to detect and track multiple maneuvering targets with time-varying motion models,an IM M-LM B Track-Before-Detect algorithm was proposed by integrating the interacting multiple models（ IM M） with LM B Track-Before-Detect filter. The sequential M onte Carlo（ SM C） implementation for IM M-LM B track-before-detect was also given. Simulation results demonstrate that the proposed algorithm can detect and track multi-targets trajectories from rawimage. In addition,it can update the model probability in an online way,which detects and tracks maneuvering targets adaptively.

关 键 词：标签多伯努利滤波器 红外搜索与跟踪 机动目标 检测前跟踪 交互式多模型 

分 类 号：TN957.51[电子电信—信号与信息处理]