战场目标识别中的D-S证据理论应用

英文篇名：Application of Dempster-Shafer evidence theory in battlefield object recognition
中文刊名：传感器与微系统
英文刊名：Transducer and Microsystem Technologies
作者：刘晓丹 ; 武俊 ; 林雪仁
英文作者：LIU Xiao-dan ; WU Jun ; LIN Xue-ren(Postgraduate Department ; Artillery Academy of PLA ; Hefei 230031 ; China)
中文关键词：地震动信号 ; 数据融合 ; 目标识别 ; Dempster-Shafer(D-S)证据推理法 ; 特征提取 ; 基本概率赋值
英文关键词：microseismic signals ; data fusion ; object recognition ; Dempster-Shafer(D-S)evidence theory ; feature extracting ; basic probability assignment
出版日期：2007-11-20
机构：解放军炮兵学院研究生系,解放军炮兵学院研究生系,解放军炮兵学院研究生系安徽合肥230031,安徽合肥230031,安徽合肥230031
年：2007
期：11
出版单位：传感器与微系统

摘要

为了解决干扰情况下地震动信号发射源的定性问题,提出了在决策层上的多传感器数据融合的识别方法。利用定位传感器组中各个传感器得到的数据,并考虑在同一个定位组中各个传感器所得数据的置信度不同,来对地震动信号发射源性质进行识别。实验结果证明:基于D-S证据理论的战场目标识别数据融合后,地震动信号发射源特性识别的可靠性明显大于单个传感器的识别效果,这也表明了多传感器融合识别的可能性和有效性。
In order to resolve problem of the recognition of ME source,the method of multi-sensor data fusion on decision level is submitd.Using the data from each sensor in a location set and considering the confidence value for each sensor,the ME source characteristics can be acquired.Experiment results show that the battlefield object recognition based on Dempster-Shafer evidence theory's data fusion result is better than that of single sensor,and the multi-sensor data fusion method is effective.

引文

[1]Dempster A P.Upper and lower probabilities induced by a multi-valued mapping[J].Annals of Mathematical Statistics,1967(38):325-339.
    [2]康耀红.数据融合理论与应用[M].西安:西安电子科技大学出版社,1997:141-148.
    [3]Selzer F,Gutfinger D.LADAR and FLIR based sensor fusion forautomatic target classification[J].SPIE(Society of Photo-OpticalInstrumentation Engineers),1988(1003):236-246.
    [4]Yair Shimshoni,Nathan Intrator.Classification of seismic signalsby integrating ensembles of neural networks[J].IEEE Transac-tions on Signal Processing,1998,46(5):1194-1201.
    [5]聂伟荣.多传感器探测与控制网络技术—地面运动目标震动信号探测与识别[D].南京:南京理工大学,2002:95-104.