Automated Platform for Microseismic Signal Analysis: Denoising, Detection, and Classification in Slope Stability Studies

Abstract

Microseismic monitoring has been increasingly used in the past two decades to illuminate (sub)surface processes such as landslides, due to its ability to record small seismic waves generated by soil movement and/or brittle behaviour of rock. Understanding the evolution of landslide processes is of paramount importance in predicting or even avoiding an imminent failure. Microseismic monitoring recordings are often continuous, noisy and consist of signals emitted by various sources. Manually detecting and distinguishing the signals emitted by an unstable slope is challenging. Research on automated end-to-end denoising, detection, and classification of microseismic events, as an early warning system, is still in its infancy. To this effect, our work is focused on jointly evaluating and developing suitable approaches for signal denoising, accurate event detection, non site-specific feature construction, feature selection and event classification. We propose an automated end-to-end system that can process big data sets of continuous seismic recordings fast and demonstrate applicability and robustness to a wide range of events (distant and local earthquakes, slidequakes, anthropogenic noise etc.). Algorithmic contributions lie in novel signal processing and analysis methods with fewer tunable parameters than the state of the art, evaluated on two field datasets and benchmarked against the state of the art.

Graph-Based Feature Weight Optimisation and Classification of Continuous Seismic Sensor Array Recordings

“…Next, similar to [ 11 ], Neyman–Pearson lemma removes the stacked signals that most likely contain only background noise with low SNR. After concatenating the remaining consecutive windows, we form new windows .…”

“…With the detected signals from continuous multi-channel recordings, it is necessary to develop interpretable and effective feature engineering for event classification. The main challenges in classifying seismic signals are: (1) lack of open access annotated datasets [ 11 ]; (2) imbalanced catalog of labeled events, caused by the sparsity of events of interest [ 11 ]; (3) high similarities between unknown natural and anthropogenic “interfering” signals and events of interest in time and/or frequency domain [ 12 ]. Feature engineering is a key step towards efficient signal classification as a large set of features with redundant information could easily increase the processing time and cause classifier overfitting, multicollinearity, and suboptimal feature ranking at the selection stage [ 2 ].…”

“…Feature engineering is a key step towards efficient signal classification as a large set of features with redundant information could easily increase the processing time and cause classifier overfitting, multicollinearity, and suboptimal feature ranking at the selection stage [ 2 ]. Feature construction for seismic events was discussed in detail in [ 11 ], where temporal, spectral and cepstral features and combinations thereof are derived from the raw denoised measurements. Feature extraction [ 4 ] and selection [ 2 ] are commonly used for dimensionality reduction of the feature space, thus decreasing required storage requirements, and testing and training times of the classifier.…”

Target Detection and Recognition of Ground Penetrating Radar using Morphological Image Analysis and Graph Laplacian Regularisation

“…2), we use morphological image analysis to construct features and Graph Laplacian Regularisation (GLR)-based label propagation, as advanced semi-supervised classifier. This has been shown to perform well when the number of labelled samples is low and dataset is noisy [11], [12].…”

“…Classification via Graph Laplacian Regularisation (GLR) has been widely used to classify image and time-series signals, especially when the number of labelled signals that can be used for training is small [10], [11]. In this paper, we use normalised GLR (identified in [12] as the best performing semi-supervised regularisation on graphs classification method for seismic signals), to identify the time slices which contain information about the real targets in order to reduce the false alarm rate and locate the real targets.…”