摘要
为进一步提高弹道导弹目标多传感器综合识别正确率,提出了一种基于二维主成分分析(Two-Dimensional Principal Component Analysis,2DPCA)的多传感器特征级综合识别方法。该方法将多个传感器的特征集经标准化后组合成二维特征矩阵,引入图像压缩技术中的2DPCA方法进行特征提取,然后将其用于弹道导弹目标特征级融合识别。以3部雷达部署下弹头目标的雷达散射截面积(Radar Cross Section,RCS)特征融合为例进行仿真验证,结果表明:相比于传统的主成分分析(Principal Component Analysis,PCA),2DPCA的识别率更高,计算复杂度更低,为弹道导弹目标识别提供了一种新的思路。
In order to further improve the multi-sensor comprehensive identification accuracy of ballistic missile targets,a multi-sensor feature-level comprehensive recognition method based on Two-Dimensional Principal Component Analysis（2DPCA） is proposed. This method combines the feature sets of multisensors into a two-dimensional feature matrix after normalization,introduces the 2DPCA method in the image compression technology for feature extraction and then applies it to the ballistic missile target feature-level fusion identification. By taking Rardar Cross Section（RCS） feature fusion of 3 radars in different deployment as an example,simulation and verification are carried out. The simulation experiment result shows that the method increases target classification accuracy and reduces the computational complexity,comparing to traditional Principal Component Analysis（PCA）. It provides a new idea for fusion identification of ballistic missile targets.
作者
李陆军
杨源
赵兴刚
潘小平
王志刚
LI Lu-jun;YANG Yuan;ZHAO Xing-gang;PAN Xiao-ping;WANG Zhi-gang(Troop No.93975 of PLA,Urumqi 830000,China;Department of Postgraduate Management,Air Force Early Warning Academy,Wuhan 430019,China;Troop No.66136 of PLA,Beijing 100043,China)
出处
《装甲兵工程学院学报》
2018年第4期57-62,共6页
Journal of Academy of Armored Force Engineering
基金
国家自然科学基金青年项目(61401503
61602506)
中国博士后基金资助项目(20110491889)
全军军事类研究生课题(2014JY545)
关键词
弹道导弹
目标识别
特征级融合
二维主成分分析(2DPCA)
二维特征矩阵
ballistic missile
target identification
feature-level fusion
Two-Dimensional Principal Com-ponent Analysis （2DPCA）
two-dimensional feature matrix
