An Adaptive Noise Reduction Method for Microseismic Data Using Parameter-Free Fuzzy Clustering

In microseismic monitoring, the collected data is distorted by noise, bringing difficulties and challenges to the subsequent processing of event identification, localization, and others. Wavelet packet transform (WPT) is one of the most commonly used noise reduction methods, but the usage of threshold can significantly lead to the informative content loss of target signals, especially at high frequency or low signal-to-noise ratio (SNR). To address above issues, this paper proposes a new noise reduction method based on WPT, supported by Fuzzy C-Means clustering (FCM) and k-nearest neighbors (KNN) algorithms. The experiment results show that the proposed method has less performance fluctuation than WPT at low frequencies, and outperforms WPT and FCM at high frequencies and low SNR.