数字化地震记录震相自动识别的方法研究

针对目前震相自动识别方法不能自动给出震相识别区间，以及不能确定识别出震相名称的问题，运用震相的运动学特征，由平均速度模型和J-B走时表数据，自动计算近震、远震和极远震的震相走时及震中距。对多尺度小波分解进行单支重构作为识别不同震相的分析信号。先求出初至震相和最大面波到时，估算出震中距，然后找到S波或PP波到时，求出准确的震中距，即可自动给出各震相的识别区间，采用线性偏振法在给定区间中识别出震相的初至时刻。由于该方法采用的是先明确要找什么震相，再由该震相的走时确定寻找区间，所以找出的初至就是要寻找的震相，自然解决了识别出震相的名称问题，从而实现了对震相的全程自动识别。

Studies on Auto-distinguishing Phase of Digital Seismic Recordings

Aiming at the problems that the current phase-distinguishing method can not give out the distinguishing interval and can not distinguish the phase name, the phase kinematic characteristics are used to automatically calculate the phase travel time of near earthquake,distant earthquake and ultra-distant earthquake and the epicentral distance from the averaging velocity model and J-B travel-time table. The reconstruction of single branch is performed to wavelet decomposition of multi-sizes, which is taken as the analysis signal of distinguishing different phases. Firstly, the first arrival phase and the arrival time of maximum surface wave are extracted and the epientral distance is evaluated. And then the arrival time of S wave or PP wave is found and precise epincentral distance is determined, namely, the distinguishing interval of each phase can be given automatically. The linear polarization method is used to distinguish the first arrival moment of the seismic phase in the given interval. Due to this method,it needs to make clear what phase should be found first and then determines the searching interval by travel time of seismic phase, so the found first arrival time of the phase is also the phase that want to find. Naturally, the phase-name distinguishability is solved, and the omnidistant automatic distinguishabitity to the seismic phase has been recognized. According to the thinking and method mentioned above, programming has recognized the automatic distinguish to the actual seismic phase. Seen from the distinguishing effects, this method is feasible and practical.