An Improved Automatic First Arrival Picking Method Based on EMD and AIC

Robust automatic first-arrival picking of is important in microseismic monitoring and acoustic emission (AE) data processing. Traditional Akaike information criterion (AIC) can be easily suffered from noise and inappropriate window size. In this paper, we propose an improved first-arrival picking approach based on the empirical mode decomposition (EMD) and AIC. AE signal or microseismic data is firstly decomposed into a series of intrinsic mode functions (IMFs) representing generally simple oscillatory modes. Then the AIC criterion is applied on one of IMFs representing the main vibration mode of the signal to estimate the arrival time. The selected IMF has similar waveform and amplitude spectrum to the noise-free signal. Both synthetic data and real acoustic emission data with different level of noise are used in the testing of this EMD+AIC method. Comparison between EMD+AIC and traditional AIC shows obvious improvement and demonstrates the feasibility of the new method. This new first-arrival picking method is robust and applicable for noisy AE data or field microseismic data with low SNR.