利用人工神经网络理论对地震信号及地震震相进行识别  

THE SEISMIC SIGNAL AND PHASE RECOGNITION BY USING ARTIFICIAL NEURAL NETWORK THEORY

作　　者：张范民[1] 李清河[1] 张元生[1] 盛国英 范兵[1] 

机构地区：[1]中国地震局兰州地震研究所,兰州730000

出　　处：《西北地震学报》1998年第4期43-49,共7页Northwestern Seismological Journal

摘　　要：根据人工神经网络理论,初步研究了人工神经网络对地震信号及震相进行识别的能力.为了进行识别,将三分向地震资料的矢量模作为神经网络的输入.结果表明:用此方法确定地震震相和到时是十分有效的,特别是对于信噪比较高的地震记录,效果更好.According to the theory of artificial neural networks,a preliminary study has been performed to test the ability of recognising seismic signals and phases by using artificial neural networks,and the phase recognition has been achieved for three-component recordings by using vector modulus of these seismic records as the network input.Results show that the method is very valid in the determination of seismic phases and arrivals,especially for high signalnoise ratio records.

关 键 词：地震信号 识别法 震相 人工神经网络 地震 

分 类 号：P315.63[天文地球—地震学]