王旭,
李智涛,
谭巧,
王建军
中国地震局地壳应力研究所, 北京 100085
基金项目: 国家自然科学基金（41631073，41874019）和中央级公益性科研院所基本科研业务专项资助（JDZ2017-19）共同资助
详细信息
作者简介: 付继华, 男, 博士, 副研究员, 主要从事精密测量与智慧系统研究.E-mail:fujh@email.eq-icd.cn
中图分类号: P315 收稿日期:2017-12-06
修回日期:2019-01-22
上线日期:2019-04-05
Automatic picking up earthquake's P waves using signal-to-noise ratio under a strong noise environment
FU JiHua,WANG Xu,
LI ZhiTao,
TAN Qiao,
WANG JianJun
Institute of Crustal Dynamics, China Earthquake Administration, Beijing 100085, China
MSC: P315
--> Received Date: 06 December 2017
Revised Date: 22 January 2019
Available Online: 05 April 2019
摘要
摘要:大数据量、强噪声环境给地震P波到时的自动提取带来很大挑战.针对此问题，本文通过构建特殊的特征函数，建立SNR与STA/LTA的内在联系，提出两种基于SNR的地震P波到时自动提取方法，即基于SNR的STA/LTA方法与基于SNR的综合方法.这两种方法分别是运用SNR概念对传统STA/LTA方法和STA/LTA与AIC综合方法的改进.仿真分析结果表明：对于弱噪声环境（10 dB）和一般噪声环境（6 dB），本文方法较传统STA/LTA方法对地震P波到时提取的准确度更高；而对于强噪声环境（3 dB），本文方法仍能准确提取地震P波到时，而传统STA/LTA方法则出现了较大的误判率（10%）与漏判率（65%）.本文方法为STA/LTA赋予了明确的物理意义，使其阈值的选取建立在严密的数学推导之上.另外，本文方法在进行地震P波到时自动提取的同时，兼具数据预处理功能，无需额外的基线校正或高通滤波，因而具有较好的实时性.
关键词: P波到时/
地震预警/
信噪比/
STA/LTA/
AIC
Abstract:Big data and strong noise environments bring great challenges to pick up P waves automatically. In order to solve this problem, a special characteristic function is constructed with the conception of Signal-to-Noise Ratios (SNR). By using this function, an internal relationship between the SNR and the Short-Term Average and Long-Term Average ratio (STA/LTA) is built. And two novel SNR-based P waves' picking methods are proposed, namely the SNR-based STA/LTA method and the SNR-based comprehensive method which are respectively the improvements of the traditional STA/LTA method and the comprehensive method of STA/LTA and Akaike Information Criteria (AIC) by using the SNR conception. The simulation analysis shows that under a weak noise circumstance (10 dB) and normal noise circumstance (6 dB) the two proposed methods have higher accuracy than the traditional STA/LTA method. And under a strong noise circumstance (3 dB), both the methods can accurately pick up the seismic P waves without any regulations, whereas the traditional STA/LTA method has a big error ratio (10%) and a large missing ratio (65%). The proposed methods give STA/LTA a clear physical meaning, and their thresholds can be obtained based on rigorous mathematical derivation. In addition, the proposed methods have favorable real-time performances because they can process data without requirement of additional baseline correction or high-pass filtering.
Key words:P wave's arrival/
Earthquake early warning/
Signal-to-noise ratios/
Short-Term Average and Long-Term Average ratio (STA/LTA)/
Akaike Information Criteria (AIC)
PDF全文下载地址:
http://www.geophy.cn/data/article/export-pdf?id=dqwlxb_14947
