Recognition criteria of earthquakes and explosions based on wavelet packet analysis

This paper uses sym5 wavelet packet basics function to decompose waveforms of small magnitude natural earthquakes and artificial explosions. Through spectrograms the differences of frequency components could be visually seen, the explosion frequency is simple and the aggregation of spectrogram is better than earthquake. Combining the wavelet packet transform results of P-wave and S-wave, the P/S energy ratio is defined so as to seek quantitative recognition indices. Analyses on the better classified P/S energy ratios show that the explosion’s energy of P mainly concentrates on the frequency band of 3.125—9.375 Hz, and the earthquake’s frequency are more complex, S are also developed at high frequency 12.5—23.4375 Hz, indicating that the differences between P-wave and S-wave of explosion are greater than natural earthquake in these frequency bands. As a supplement of wavelet packet criteria, we also extract and analyze the energy ratios of P and S waves. The results of energy ratio criteria show that earthquake and explosion have different frequency component and we can extract effective recognition criteria by wavelet packet transform.