Wavelet Transform-Based Fuzzy Clustering Microseismic First-Arrival Picking Method

Microseismic arrival time picking serves as the foundation for microseismic source localization and holds significant importance in the field of microseismic monitoring. Traditional methods, such as Short-Time Average/Long-Time Average (STA/LTA) and clustering methods based on STA/LTA as feature vectors, require manual adjustments of the time window parameters to achieve accurate picking. Furthermore, they are susceptible to inaccuracies in high-background noise environments. And in response to these challenges, this study introduces a fuzzy clustering algorithm based on Continuous Wavelet Transform (CWT-FCM) for microseismic arrival time picking. This method begins by transforming raw data into the wavelet domain and selecting scales with relatively large standard deviations as input for the fuzzy clustering process. Ultimately, it identifies the initial arrivals of microseismic events within the resulting clusters. In this study, our proposed method is applied to microseismic datasets with low signal-to-noise ratios as well as real data, successfully and accurately picking microseismic arrivals. Compared with traditional methods, our approach demonstrates increased robustness and practical value in high-interference scenarios. Notably, it eliminates the need for manual parameter adjustments, thereby enhancing efficiency and precision in automated microseismic signal picking and establishing a foundational dataset for subsequent automatic and high-precision microseismic arrival time localization.