Explainable time–frequency convolutional neural network for microseismic waveform classification

Bi, Xin; Zhang, Chao; He, Yao; Zhao, Xibin; Sun, Yongjiao; Ma, Yuliang

doi:10.1016/j.ins.2020.08.109

Cited by 33 publications

(12 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…XTF-CNN Uses frequency and time domain series as the features of the Convolutional Neural Network. This method also has a subsystem to map it back as explainable traces [9].…”

Section: Ann Identificationmentioning

confidence: 99%

“…There will be a change in the network architecture, such as the number and the type of layers [8]. Then, the proposed system will include a subsystem in one of the micro-seismic mining systems [9], which has included the frequency domain of the signal as a feature of the neural network. Also there will be additional pre-processing data, a dataset balancing method, the SMOTE, to further improve the performance.…”

Section: Cpicmentioning

confidence: 99%

“…After a standard filtering process to increase SNR, the signal will be transformed to the frequency domain for classifier input features. The discrete cosine transform is chosen because this system only needs the real part of the signal [9]. Formula (1) was used for this research implementation.…”

Section: Discrete Cosine Transformmentioning

confidence: 99%

“…Frequency Transformation Several windows of the filtered signal. Discrete Cosine Transformation [9]. Discrete Frequency domain of the filtered signal.…”

Section: Filteringmentioning

confidence: 99%

See 3 more Smart Citations

P-Wave detection using deep learning in time and frequency domain for imbalanced dataset

Sugondo

Machbub

2021

Heliyon

View full text Add to dashboard Cite

Convolutional neural networks with time and frequency domain can detect Earthquake Primary Wave (P-Wave) signals well.• The Synthetic Minority Oversampling Technique (SMOTE) can increase the imbalanced earthquake dataset training performances. • P-Wave detection with several time windows and classifiers can increase overall performance, surpassing traditional method. • The proposed system can achieve a 0.17second execution time with 0.4 seconds periodic time.

show abstract

“…XTF-CNN Uses frequency and time domain series as the features of the Convolutional Neural Network. This method also has a subsystem to map it back as explainable traces [9].…”

Section: Ann Identificationmentioning

confidence: 99%

Section: Cpicmentioning

confidence: 99%

Section: Discrete Cosine Transformmentioning

confidence: 99%

“…Frequency Transformation Several windows of the filtered signal. Discrete Cosine Transformation [9]. Discrete Frequency domain of the filtered signal.…”

Section: Filteringmentioning

confidence: 99%

See 2 more Smart Citations

P-Wave detection using deep learning in time and frequency domain for imbalanced dataset

Sugondo

Machbub

2021

Heliyon

View full text Add to dashboard Cite

show abstract

“…With the rapid development in the field of computers, artificial intelligence technology has been widely used in seismic/microseismic processing and disaster prediction [15][16][17][18]. Xin et al (2021) [19] proposed an explainable time-frequency convolutional neural network (CNN) to provide an excellent classification performance and explainability. Liang et al (2021) [20] combined multiple base learners and classifiers to estimate the probability of short-term rockburst risks and achieved good performance.…”

Section: Introductionmentioning

confidence: 99%

Multi-Classification of Complex Microseismic Waveforms Using Convolutional Neural Network: A Case Study in Tunnel Engineering

Zhang

Zeng

et al. 2021

Sensors

View full text Add to dashboard Cite

Due to the complexity of the various waveforms of microseismic data, there are high requirements on the automatic multi-classification of such data; an accurate classification is conducive for further signal processing and stability analysis of surrounding rock masses. In this study, a microseismic multi-classification (MMC) model is proposed based on the short time Fourier transform (STFT) technology and convolutional neural network (CNN). The real and imaginary parts of the coefficients of microseismic data are inputted to the proposed model to generate three classes of targets. Compared with existing methods, the MMC has an optimal performance in multi-classification of microseismic data in terms of Precision, Recall, and F1-score, even when the waveform of a microseismic signal is similar to that of some special noise. Moreover, semisynthetic data constructed by clean microseismic data and noise are used to prove the low sensitivity of the MMC to noise. Microseismic data recorded under different geological conditions are also tested to prove the generality of the model, and a microseismic signal with Mw ≥ 0.2 can be detected with a high accuracy. The proposed method has great potential to be extended to the study of exploration seismology and earthquakes.

show abstract

Strong noise-tolerant deep learning network for automatic microseismic events classification

Jian

Tuo

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

Explainable time–frequency convolutional neural network for microseismic waveform classification

Cited by 33 publications

References 20 publications

P-Wave detection using deep learning in time and frequency domain for imbalanced dataset

P-Wave detection using deep learning in time and frequency domain for imbalanced dataset

Multi-Classification of Complex Microseismic Waveforms Using Convolutional Neural Network: A Case Study in Tunnel Engineering

Strong noise-tolerant deep learning network for automatic microseismic events classification

Contact Info

Product

Resources

About