图像目标检测前跟踪的广义多伯努利滤波算法
Received:April 11, 2017  Revised:August 21, 2017  点此下载全文
引用本文:
Hits: 285
Download times: 
Author NameAffiliationE-mail
Shi zhiguang ATR Lab,NUDT,Hunan Changsha. 410073 szgstone75@sina.com 
Zhou jianxiong ATR Lab,NUDT,Hunan Changsha. 410073  
Zhang yan ATR Lab,NUDT,Hunan Changsha. 410073  
中文摘要:针对目标影响区域重叠时的图像目标检测前跟踪问题,推导了基于多伯努利滤波器的多目标联合检测与跟踪算法。该方法在分析多个目标叠加条件下的观测似然函数基础上,利用预测得到的目标状态对观测似然函数进行估计,从而消除目标叠加对观测更新带来的影响。该方法在目标预测与跟踪阶段皆保持了目标状态的多伯努利分布特性,是较为严格意义上的多伯努利多目标滤波器,可应用于一般图像观测条件下(目标重叠或非重叠)的目标检测前跟踪。我们给出了该算法的实现步骤,并通过加标签的方法,更准确地实现目标轨迹提取和虚假目标剔除,最后通过计算机仿真实验验证了所提算法的有效性。
中文关键词:多伯努利滤波器,图像,跟踪,检测前跟踪,重叠目标.
 
Generalized Multi-Bernoulli Filters for Track-before-detect of Objects from Image Observations
Abstract:In this paper, we propose a Generalized Multi-Bernoulli Filter for Track-before-detect (GMB - TBD) of objects from image observations when the objects’ influence region overlapping. We analysis the overlapping objects’ measurement likelihood function, estimate this likelihood function by predicted objects’ states and eliminate the objects’ overlapping influence on object’s states updating using this estimation. In this filter, the predicted and updated objects’ states are strictly assumed as Multi-Bernoulli RFS, so it’s a truly Multi-Bernoulli based TBD filter and it can be used under both the objects’ influence region overlapping and non-overlapping situations. We give the filter’s realization steps, prune and extract the objects’ tracks by labeling the Multi-Bernoulli components. Lastly, we test the GMB-TBD filter’s performance by computer Monte-Carlo simulations.
keywords:Multi-Bernoulli filter, images, tracking, track before detect (TBD), overlapping objects.
  View/Add Comment  Download reader

Copyright:《Journal of Infrared And Millimeter Waves》