2016年发布APASVO-p波震相自动拾取分析

• Why automatic attractive?

1. large amount of seismic data ;
2. if manually,it depends om experience of analyst;
3. Quliaty can be obscured by several factors such as background and non-stationary noise from diverse sources;

• 该软件使用的方法为

       STA-LTA（针对振幅突变识别很有效）；

       AMPA；

       An autogressive AIC-based picking method ;

• 软件优点

1. 适用于多平台；
2. 可以处理大批量，不同文件格式的数据；
3. 可用于microseismicity的震相拾取；

• 软件相关依赖

       Scientific  computing packages such as ObsPy,SciPy,NumPy,and Matplotlib

• 拾取精度对后续工作的影响

1.  tomographic inversion of active-source data(震相成像分析)
2. HYPOCENTER LOCATION

• 作者思路：将前人拾取方法分为用于粗糙拾取（ relatively imprecise）和精确拾取（precise）两种

1. STA／LTA：naive version of the method has too broad range in parameter configuration;(可能需要大量统计实验对于不同距离范围的地震)
2. direct use of an envelope function with a dynamic trigger level ;(Baer and Kradolfer,1987)
3. higher-order statistic;(Saragiotis et al.,2002;Kuperkoch et al,.2010;Nippress et al.,2008)
4. polarization approaches;(Montalbetti and Kanasewich,1970;Kurzon et al.,2014)
5. local-maxima distribution(Panagiotakis et al.,2008)
6. wavelet-based methods(Zhang et al.,2003)
7. even pattern recognition methods(Tong,1995)
8. neural networks(Dai andMacBeth,1995;Gentili and Michelini,2006)
9. computer vision inspired methods(Joswig,1990)
10. Autogressive models:the standard one named AR-AIC  (建立背景噪声和到来震相的自回归模型)：find the phase coming point that the locally stationary AR model attains the minimum value of the Akaike information criterion.(达到某个标准的最小值时)；：该方法拾取较精确，但由于计算比较强烈建议用于已经有过初步估计的初动拾取工作中；

• 解决信噪比低的信号的方案

AMPA  algorithm（Adaptive Muti-Band Picking Algorithm）:related article-Alvarez et al.,2013

It  applied a serises of filtering stages to the input signal in order to 减弱环境噪声并加强震相的到来。

1. step1:using filter bank to eatimate the signal envelope and equalize (补偿) noise for each subband.filter band
2. the resultant  envelope is filterd by a use of set of filters,each one designed different length in order to enhance the phase arrival and lower emergent  noised.

• 软件方法支撑

STA／LTA 和AMPA  picking algorithm 为主 ；AIC为辅；

• 软件框架

1. apasvo-gui:交互型界面，允许用户手动改值；
2. apasvo-detector:允许使用命令行来计算；有批处理和监督模式两种选择模式，监督用来交互型地判断是否是满意的拾取，是否接受。批处理则在处理过程中不过问。
3. apasvo-generator:允许使用命令行来产生地震波轨迹，并允许其中有背景噪声；