The Use of Wavelet-Based Denoising Techniques to Enhance the First-Arrival Picking on Seismic Traces

Arrival time picking is a critical step in the analysis of geophysical data. This study aims at improving the standard short-term-average/long-term-average (STA/LTA) method using soft-thresholding denoising techniques based on the discrete wavelet transform. The suggested method has been first tested on simulated seismic traces. For the purpose of choosing the best parameters, we have investigated different analyzing wavelets, threshold selection rules (“minimaxi”, “universal”, “rigorousSURE”, and “heuristicSURE”), and threshold rescaling types (global and adaptive). It is shown that the best results are obtained using the Haar wavelet, the 'universal' rule, and the multiple-scale-dependent rescaling type. Afterward, simulated data sets with varying signal-to-noise ratio (SNR) have been used to compare the wave onset times picked using both STA/LTA versions (standard and denoised) and the modified energy ratio (MER) algorithm. The results show that, for data sets with high SNR values (greater than three), the MER algorithm yields the most accurate arrival times whereas, for low SNR values varying from 3 to 1.5, the denoised STA/LTA algorithm is the most effective picking algorithm. Furthermore, the picking techniques have been applied on real seismic traces recorded in the Algerian Sahara. It is again confirmed that the proposed technique provides the most reliable picking for high-noise traces. To conclude, the denoised STA/LTA algorithm is a powerful tool for identifying the first arrival for high-noise signals with SNR values lower than three and can tolerate an SNR value of about 1.5.