Source Separation on Seismic Data: Application in a Geophysical Setting

Moni, Aishwarya; Bean, Christopher J.; Lokmer, Ivan; Rickard, Scott

doi:10.1109/msp.2012.2184229

Cited by 10 publications

(5 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However recent works have proposed several more sophisticated schemes to “clean” seismic traces. These include methods based on the independent component analysis (Cabras et al., 2010; Comon, 1994; Moni et al., 2012), beamforming methods (Boué et al., 2013; Brooks et al., 2009; Gibbons et al., 2008), and MUltiple SIgnal Classification (Bear et al., 1999; Schmidt, 1986). All of these methods, however, can fall short when the noise shares frequencies with the earthquake generated signal.…”

Section: Introductionmentioning

confidence: 99%

Cold Diffusion Model for Seismic Denoising

Trappolini,

Laurenti,

Poggiali

et al. 2024

Journal of Geophysical Research: Machine Learning and Computati

View full text Add to dashboard Cite

Seismic waves contain information about the earthquake source, the geologic structure they traverse, and many forms of noise. Separating the noise from the earthquake is a difficult task because optimal parameters for filtering noise typically vary with time and, if chosen inappropriately, may strongly alter the original seismic waveform. Diffusion models based on Deep Learning have demonstrated remarkable capabilities in restoring images and audio signals. However, those models assume a Gaussian distribution of noise, which is not the case for typical seismic noise. Motivated by the effectiveness of “cold” diffusion models in speech enhancement, medical anomaly detection, and image restoration, we present a cold variant for seismic data restoration. We describe the first Cold Diffusion Model for Seismic Denoising (CDiffSD), including key design aspects, model architecture, and noise handling. Using metrics to quantify the performance of CDiffSD models compared to previous works, we demonstrate that it provides a new standard in performance. CDiffSD significantly improved the Signal to Noise Ratio by about 18% compared to previous models. It also enhanced Cross‐correlation by 6%, showing a better match between denoised and original signals. Moreover, testing revealed a 50% increase in the recall of P‐wave picks for seismic picking. Our work show that CDiffSD outperforms existing benchmarks, further underscoring its effectiveness in seismic data denoising and analysis. Additionally, the versatility of this model suggests its potential applicability across a range of tasks and domains, such as GNSS, Lab Acoustic Emission, and Distributed Acoustic Sensing data, offering promising avenues for further utilization.

show abstract

Section: Introductionmentioning

confidence: 99%

Cold Diffusion Model for Seismic Denoising

Trappolini,

Laurenti,

Poggiali

et al. 2024

Journal of Geophysical Research: Machine Learning and Computati

View full text Add to dashboard Cite

show abstract

“…Therefore, the seismovolcanic monitoring (similarly to many other areas) must rely more on data‐intensive automatic methods for analysis and classification of signals leading to the idea of applying methods from the area of machine learning (e.g., Carniel, , ; Orozco‐Alzate et al., ). Machine learning approaches related to seismovolcanic data are most of the time applied on single time series from individual stations to perform blind source separation with dimension reduction methods such as Independant Component Analysis (Acernese et al, ; Cabras et al, , ; Ciaramella et al, ; Capuano et al, ), Nonnegative Matrix Factorization (Cabras et al, , ), or Degenerate Unmixing Estimation Technique (Moni et al, ). A few machine learning studies of seismovolcanic signals adopt a multistation approach but dealing with features (a machine learning terminology that stands for any attribute of the waveform) derived from individual stations of the network (e.g., waveform, spectrum, spectrogram,…).…”

Section: Introductionmentioning

confidence: 99%

Network‐Based Detection and Classification of Seismovolcanic Tremors: Example From the Klyuchevskoy Volcanic Group in Kamchatka

et al. 2018

View full text Add to dashboard Cite

We develop a network‐based method for detecting and classifying seismovolcanic tremors. The proposed approach exploits the coherence of tremor signals across the network that is estimated from the array covariance matrix. The method is applied to four and a half years of continuous seismic data recorded by 19 permanent seismic stations in the vicinity of the Klyuchevskoy volcanic group in Kamchatka (Russia), where five volcanoes were erupting during the considered time period. We compute and analyze daily covariance matrices together with their eigenvalues and eigenvectors. As a first step, most coherent signals corresponding to dominating tremor sources are detected based on the width of the covariance matrix eigenvalues distribution. Thus, volcanic tremors of the two volcanoes known as most active during the considered period, Klyuchevskoy and Tolbachik, are efficiently detected. As a next step, we consider the daily array covariance matrix's first eigenvector. Our main hypothesis is that these eigenvectors represent the principal components of the daily seismic wavefield and, for days with tremor activity, characterize dominant tremor sources. Those daily first eigenvectors, which can be used as network‐based fingerprints of tremor sources, are then grouped into clusters using correlation coefficient as a measure of the vector similarity. As a result, we identify seven clusters associated with different periods of activity of four volcanoes: Tolbachik, Klyuchevskoy, Shiveluch, and Kizimen. The developed method does not require a priori knowledge and is fully automatic; and the database of the network‐based tremor fingerprints can be continuously enriched with newly available data.

show abstract

“…Parra & Spence 2000;Vincent et al 2006) but also in geophysics (e.g. Ikelle 2007;Moni et al 2012;Takahata et al 2012;Liu & Dragoset 2013). In our context, the purpose of BSS is to unmix mixtures of signals, retrieving the isolated signals (mixing components), and the amplitude with which they contribute to the mixture (mixing amplitudes).…”

Section: Introductionmentioning

confidence: 99%

“…BSS can be used to separate signals with different slowness from these mixtures. Based on this idea, a BSS method called DUET has been used in different settings to unmix signals from the volcano Mt Etna (Moni et al 2012), and also from secondary microseisms (Moni et al 2013).…”

Section: Introductionmentioning

confidence: 99%

Blind source separation of temporally independent microseisms

Meschede

Stutzmann

Schimmel

2018

Geophysical Journal International

View full text Add to dashboard Cite

Secondary microseismic sources emit seismic waves over long time spans. Reoccurring signals with similar slowness and frequency therefore arrive at seismic arrays. Blind source separation techniques can be used to identify and isolate such reoccurring signals from other signals and from diffuse seismic noise. Along these lines, we use non-negative matrix factorization as blind source separation technique to decompose continuous seismic array records. We model the recorded energy as a mixture of a few components with static slowness-frequency and time-dependent amplitudes. Components and amplitudes are fitted to optimally explain the recorded seismic energy over time. These components represent secondary microseismic signals with quasi-static slowness-frequency vector and fluctuating amplitude. Each fitted component reveals the geographical origin (through the slowness-frequency vector) and time evolution of an active secondary microseism with high precision because it is separated from other signals and diffuse seismic noise. Furthermore, relative traveltimes can be automatically extracted for the signals that correspond to a specific component that can potentially be used in tomographic studies. We show two examples of seismic signals that were extracted with this technique, one focusing on P waves from the typhoons Goni and Atsani, and the other showing secondary microseism PKP signals from typhoon Glenda.

show abstract

Source Separation on Seismic Data: Application in a Geophysical Setting

Cited by 10 publications

References 11 publications

Cold Diffusion Model for Seismic Denoising

Cold Diffusion Model for Seismic Denoising

Network‐Based Detection and Classification of Seismovolcanic Tremors: Example From the Klyuchevskoy Volcanic Group in Kamchatka

Blind source separation of temporally independent microseisms

Contact Info

Product

Resources

About