A study on noise reduction and automatic P-phase onset time picking technology of weak micro-seismic signals from underground mines

Abstract

The frequency range of the microseismic signal is wide, with contamination of noise including human disturbance, power frequency interference and other signals from underground mines, which greatly reduce the signal to noise ratio and raise a great challenge to accurate detection and picking of the P-phase onset time in noisy micro-seismic data from underground mines. Reliable P phase onset time picking plays a key role in accurate source location. Facing this situation, in this paper, synchrosqueezing wavelet transform (SST) is used to filter the micro-seismic signal in time-frequency domain, and the short-term to long-term average (STA/LTA) ratio combined with AIC method is used to finish P phase onset time picking. The objectives of this method are to improve the signal to noise ratio and provide an automatic means of P phase onset time picking. Firstly, the synchronous squeeze wavelet transform is used to filter the signal over several scales, and then the automatic P-phase onset is achieved by using the short-term to long-term average (STA/LTA) ratio combined with AIC method. Finally, the actual mine data are processed in this method, and the result of automatic P-phase onset time picking is compared with manual picking. The results show that the method provides a more reliable means of the P phase onset arrival and accurate matching the manual picking for low signal to noise ratio data.