摘要
利用2009年夏季青海地区的快电场变化测量仪的野外观测资料,对120例地闪和77例云闪辐射场信号的分形特征进行了深入研究,结果表明地闪辐射场信号的分形维数与云闪辐射场信号的分形维数有明显的差别,再利用闪电的分形维特征,构造了5个特征值,将其作为支持向量机的输入进行地闪和云闪不同放电类型的识别,有效识别率达到95%以上;通过构造地闪辐射场时间序列信号的分形维数轨迹图表明分形维数最低点对应于原时间序列的回击位置,利用分形维数轨迹中的最低点的位置能够快速准确地对地闪辐射场信号的回击点进行检测,检测率可达到100%.分形维是闪电的一种具有鉴别性的特征,可用于闪电的智能分析与自动化处理.
By analyzing the fractal feature of 120 cloud-to-ground lightning signals and 77 intracloud lightning signals obtained by the fast antenna system in Qinghai area during the summer of 2009, the results show fractal dimension of cloud-to-ground lightning signal is obviously different from that of intracloud lightning signal. Then 5 characteristic values of fractal dimension are used to recognize the discharge types of lightning signal via support vector machine, and the recognition rate is higher than 95%. The construction of cloud- to-ground lightning time series signal fractal dimension trajectory map shows that the fractal dimension minimum value corresponds to the return stroke of the original time series signal, which can be used to quickly and accurately detect the return stroke of lightning signal, and the detection rate can reach 100%.The fractal dimension is a discriminatively physical property which can be used for intelligently analyzing and automatically processing the lightning signal.
出处
《物理学报》
SCIE
EI
CAS
CSCD
北大核心
2013年第5期566-574,共9页
Acta Physica Sinica
基金
国家自然科学基金(批准号:41075002
40775004)
国家自然科学基金重点项目(批准号:41030960)
公益性行业科研专项基金(批准号:GYHY201006005-03)资助的课题~~
关键词
闪电信号
分形维数
支持向量机
自动识别
lightning signal, fractal dimension, support vector machine, automatic recognition
