Distributed fiber optic acoustic signal denoising and P-wave arrival picking method based on POA-LMD-LSTM

To address the common issues of low signal-to-noise ratio in distributed fiber optic acoustic signals and instability in P-wave first arrival picking using the Akaike information criterion (AIC) method, a new signal denoising and P-wave first arrival picking method is proposed. This method leverages, local mean decomposition optimized by the Pelican algorithm and long short-term memory network (POA-LMD-LSTM) to predict the signal, and improves AIC method. Laboratory experimental tests indicate that the proposed method offers significant advantages over existing technologies in terms of improving the SNR of P-wave arrivals, picking stability, and accuracy. The signal-to-noise ratio of the denoised signal increases by an average factor of 1.15, and the average picking error of the P-wave arrival using the improved AIC method is only 1.0 sample point, which is significantly lower than that of the S/L-Skewness method.