首页 | 官方网站   微博 | 高级检索  
     

核爆与雷电电磁脉冲识别
引用本文:祁树锋,李夕海,韩绍卿,牛超,冯军,刘代志.核爆与雷电电磁脉冲识别[J].强激光与粒子束,2012,24(12):2935-2940.
作者姓名:祁树锋  李夕海  韩绍卿  牛超  冯军  刘代志
作者单位:1.第二炮兵工程学院, 西安 71 0025
基金项目:国家自然科学基金项目(40974037); 中国博士后科学基金项目(201150M1554)
摘    要:用经验模态分解和分形分析相结合的方法对核爆和雷电电磁脉冲信号进行了识别研究。计算了核爆和雷电电磁脉冲原始信号的盒维数,以及原始信号经验模态分解后前4阶固有模态分量(IMF)的盒维数。用最近邻法对核爆和雷电进行了识别,实验结果表明:原始信号经验模态分解后一阶、二阶IMF盒维数的识别率要略高于三阶、四阶IMF盒维数的识别率;原始信号盒维数的识别率高于经验模态分解后各阶IMF盒维数的识别率;二维和三维特征的识别率要高于一维特征的识别率,另外二维和三维特征的识别率更加趋于稳定,并且三维特征的识别率都在90%以上。

关 键 词:核爆电磁脉冲    雷电电磁脉冲    识别    经验模态分解    分形
收稿时间:2012/2/26

Discrimination of nuclear-explosion and lightning electromagnetic pulse
Affiliation:1.Second Artillery Engineering Institute,Xi’an 710025,China
Abstract:The discrimination of nuclear-explosion and lightning electromagnetic pulses was studied using empirical mode decomposition and the fractal analytical method. The box dimensions of nuclear-explosion and lightning electromagnetic pulses’ original signals were calculated, and the box dimensions of the intrinsic mode functions (IMFs) of nuclear-explosion and lightning electromagnetic pulses’ original signals after empirical mode decomposition were also obtained. The discrimination of nuclear explosion and lightning was studied using the nearest neighbor pattern classification. The experimental results show that, the discrimination rate of the box dimension based on the first and second IMF after the original signal empirical mode decomposition is higher than that based on the third and forth IMF; the discrimination rate of the box dimension based on the original signal is higher than that based on any IMF; and the discrimination rate based on two-dimensional and three-dimensional characters is higher and more stable than that based on one-dimensional character, besides, the discrimination rate based on three-dimensional character is over 90%.
Keywords:nuclear electromagnetic pulse  lightning electromagnetic pulse  discrimination  empirical mode decomposition  fractal
点击此处可从《强激光与粒子束》浏览原始摘要信息
点击此处可从《强激光与粒子束》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司    京ICP备09084417号-23

京公网安备 11010802026262号