主成分分析在震动信号目标识别算法中的应用

英文篇名：Application of principal component analysis in target recognition algorithm of seismic signals
中文刊名：华中科技大学学报(自然科学版)
英文刊名：Journal of Huazhong University of Science and Technology(Natural Science Edition)
作者：鲍必赛 ; 楼晓俊 ; 李隽颖 ; 刘海涛
英文作者：Bao Bisai1 Lou Xiaojun1 Li Junying1 Liu Haitao1 ; 2(1Key Lab of Wireless Sensor Network and Communication ; Shanghai Institute of Micro-system and Information Technology ; Chinese Academy of Science ; Shanghai 200050 ; China ; 2Wuxi SensingNet Industrialization Research Institute ; Wuxi 214135 ; Jiangsu China)
中文关键词：目标识别 ; 识别算法 ; 主成分分析 ; 震动信号 ; 特征提取
英文关键词：target recognition ; recognition algorithm ; principal component analysis ; seismic signals ; feature extraction
出版日期：2012-07-23
机构：中国科学院上海微系统与信息技术研究所无线传感器网络与通信重点实验室;无锡物联网产业研究院;
年：2012
期：07
出版单位：华中科技大学学报(自然科学版)

摘要

为了改进基于震动信号的地面运动目标识别算法,提出了一种基于主成分分析(PCA)的2次特征提取算法.首先对地面运动目标引起的震动信号进行目标特性分析,提取多维的特征值;然后利用主成分分析方法对众多的特征值进行分析,去除特征值之间的相关性,提取综合特征值并应用于分类器,得到目标识别结果.基于实地采集的地面运动目标的震动信号进行实验,结果表明:该方法有效地减少了特征值的维数和相关性,降低了分类器训练的难度和训练时间,同时提高了目标的正确识别率.
In order to improve the algorithm of ground moving targets based on seismic signals,an algorithm of second feature extraction based on principal component analysis(PCA)was presented.First the target characteristics of seismic signals caused by ground moving targets were analyzed and multi-dimensional feature vectors were extracted.Then the large number of feature vectors was analyzed through principal component analysis.After the correlation between the feature vector was removed,the integrated feature vector was extracted and used in classifier to obtain result of target recognition.Based on real seismic signals of ground targets,experiment results indicate that this method can effectively decrease the dimension and correlation of feature vectors,reduce the difficulty and classifier training time,and improve the performance of classification,providing an idea for target recognition of seismic signals.

引文

[1]Tong Libiao,Lu Wenjun,Qi Dening,et al.Target recognition based on seismic sensors and neural net-work[C]∥The Eighth International Conference on Electronic Measurement and Instruments.Xi′an:ICEMI′,2007:18-21.
    [2]黄琦.智能传感器侦察网络中的地面目标识别算法研究[D].合肥:中国科学技术大学物理学院,2006.
    [3]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:南京理工大学机械学院,2002.
    [4]聂伟荣,朱继南,郭亚军,等.地震动信号的分析与目标识别[J].电子科技大学学报,2003,32(1):26-30.
    [5]曹红兵,魏建明,刘海涛.无线传感网中多传感器特征融合算法研究[J].电子与信息学报,2010,32(1):166-171.
    [6]陈建宏,刘浪,周智勇,等.基于主成分分析与神经网络的采矿方法优选[J].中南大学学报:自然科学版,2010,41(5):1967-1972.
    [7]邢怀飞.智能光纤传感器网络地面目标识别若干问题研究[D].北京:中国科学院半导体研究所,2008.
    [8]李舜酩,李香莲.振动信号的现在分析技术与应用[M].北京:国防工业出版社,2008.
    [9]Hargrove L J,Guanglin Li,Englehart K B,et al.Principal components analysis preprocessing for im-proved classification accuracies in pattern-recognition-based myoelectric control[J].Biomedical Engineer-ing,2009,56(5):1407-1414.
    [10]Panagakis Y,Kotropoulos C,Arce G R.Non-nega-tive Multilinear Principal Component Analysis of Auditory Temporal Modulations for Music Genre Classification[J].Audio,Speech,and Language Processing,2010,18(3):576-588.
    [11]Qian Du,Wei Wei,May D,et al.Noise-adjusted principal component analysis for buried radioactive target detection and classification[J].Nuclear Sci-ence,2010,57(6):3760-3767.