一种基于PDAF的目标跟踪算法  

A Target Tracking Method Based on PDAF

作　　者：张剑波 高貂林 刘晓林 

机构地区：[1]海鹰企业集团有限责任公司,江苏无锡214061

出　　处：《海洋测绘》2016年第3期60-62,71,共4页Hydrographic Surveying and Charting

摘　　要：对目标进行跟踪的目的是通过相关和滤波处理建立目标的运动轨迹。在某型声纳多次试验中,发现采用最近邻域标准滤波器(NNSF)滤波方法时,目标跟踪不稳,跟偏、丢失等现象频现,因此对现有目标跟踪模型和算法进行研究与改进。采用基于概率数据互联滤波器(PDAF)的跟踪算法来实现对目标的跟踪。首先,使用滑窗法对目标进行跟踪起始,然后再采用概率数据互联滤波器(PDAF)进行目标关联,以保持对目标的连续跟踪。经过仿真与在某型声纳的多次试验对比可得:该关联方法对目标的跟踪结果较采用NNSF滤波方法对目标的跟踪结果有明显改进。其跟踪精度有了明显的提高,而且其跟偏、丢失的次数明显减少。The purpose of tracking target is to establish target trajectory through relevant and filter processing. In many experiments of a certain sonar,we found target tracking is unstable,offset and lost frequently,therefore the existing target tracking model and algorithm are studied and improved. The target tracking based on PDAF is realized. Firstly,using the sliding windowmethod for target tracking initiation,and then conducting target association by using PDAF to keep continuous target tracking. Through the simulation compared with many experiments of the sonar,it is available that target tracking results of the association algorithm have obvious improvement than that of NNSF filtering method. The tracking accuracy is improved,and the tracking deviation and loss are reduced.

关 键 词：声纳 目标跟踪 概率数据互联滤波器 最近邻域标准滤波器 仿真 

分 类 号：P234.2[天文地球—摄影测量与遥感]