CN112487952A - Mine microseismic signal automatic identification method based on deep learning - Google Patents
Mine microseismic signal automatic identification method based on deep learning Download PDFInfo
- Publication number
- CN112487952A CN112487952A CN202011357051.1A CN202011357051A CN112487952A CN 112487952 A CN112487952 A CN 112487952A CN 202011357051 A CN202011357051 A CN 202011357051A CN 112487952 A CN112487952 A CN 112487952A
- Authority
- CN
- China
- Prior art keywords
- signal
- signals
- waveform
- blasting
- microseismic
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
- G06F2218/10—Feature extraction by analysing the shape of a waveform, e.g. extracting parameters relating to peaks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Signal Processing (AREA)
- Probability & Statistics with Applications (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
- Geophysics And Detection Of Objects (AREA)
Abstract
The invention provides a mine micro-seismic signal automatic identification method based on deep learning, which is characterized in that a waveform database containing micro-seismic signals, blasting signals and mechanical vibration signals is established through a manual method; preprocessing waveform data in a waveform library, and enabling the three types of waveforms to have the same length by a down-sampling method; randomly selecting 70% of data of various waveforms as training samples; carrying out short-time Fourier transform on the signal waveform to generate a two-dimensional RGB picture, and respectively giving label slight shock, blasting and mechanical vibration to the RGB pictures of different waveform types; designing a convolutional neural network structure for identifying the microseismic signal, and carrying out classification and identification operations; and taking data except the training sample in the waveform library as a test sample, carrying out classification model inspection, and evaluating the recognition effect of the signal. By adopting the method, the high-quality microseismic signal identification effect with higher accuracy can be obtained, and the microseismic signal can be accurately identified from the monitoring signal of the complex mine underground environment.
Description
Technical Field
The invention relates to the technical field of mine micro-seismic signal processing, in particular to a mine micro-seismic signal automatic identification method based on deep learning.
Background
The microseismic monitoring system is generally used for ground pressure monitoring of safe mining of mines in mine production, and the sensors are arranged near a mining area and have poor arrangement conditions. Mine underground mining activities are more, generated signals have the characteristic of diversity, some non-rupture signals are similar to microseismic signals and can be captured by a microseismic monitoring system and mixed with the microseismic signals, and visual monitoring data are difficult to provide for field production service.
Although the existing microseismic monitoring systems, such as the ESG monitoring system of the Engineering seismic Group of canada (Engineering seismic Group) and the IMS monitoring system of the Institute of mine seismic in south africa (Institute of mine seismic), can set certain parameters, such as frequency and amplitude, to pre-process the triggered signal. However, the complex environment in the well causes the parameters of the noise signal to change greatly, the frequency and amplitude range is wide, and the non-microseismic signal is difficult to remove through the threshold value given by the system. In addition, poor parameter settings can filter out some of the effective microseismic signals. Therefore, it is necessary to identify the waveform when processing the microseismic signal.
Through the analysis and discovery of underground activities and monitoring signals of a mine, underground monitoring signals can be roughly divided into three categories: a rock fracture microseismic signal, a blasting signal and a mechanical vibration signal. Although the three types of signals mixed in the monitoring signal have differences in amplitude and frequency, there are many commonalities. Therefore, it is difficult to directly identify the microseismic signal from amplitude or frequency. Meanwhile, as the transmission path of the micro-seismic wave is complex, the phenomena of reflection, diffraction, attenuation and the like occur when the wave form is transmitted in a medium, and the mutual interference among the wave forms is large, the complexity of the wave form is determined, and the micro-seismic signal is more difficult to distinguish and identify.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a mine microseismic signal automatic identification method based on deep learning, which comprises the following steps:
step 1: establishing a waveform database containing microseismic signals, blasting signals and mechanical vibration signals;
step 2: processing all waveform data in a waveform database into the same waveform length by adopting a downsampling method to obtain a preprocessed signal waveform;
and step 3: generating two-dimensional RGB pictures by adopting short-time Fourier transform on the preprocessed signal waveforms, and taking all the RGB pictures as a sample set;
and 4, step 4: dividing a sample set into a training set and a verification set according to a preset proportion;
and 5: determining the number of the convolutional layers, the fully-connected layers and the number of internal neurons by using a trial-and-error method, and constructing a convolutional neural network model;
step 6: taking the RGB picture corresponding to each waveform in the training set as input, and endowing each RGB picture with a label during input;
and 7: training the convolutional neural network model through a training set to obtain a trained model;
and 8: taking the RGB pictures corresponding to each waveform in the verification set as input, endowing each RGB picture with a label when inputting, and verifying the trained model through the verification set to obtain the accuracy rate of the model for signal recognition;
and step 9: the accuracy is more than or equal to a set threshold value epsilon2The model is used as a parameter optimal model, and the parameter optimal model is used for automatically identifying the microseismic signals of the monitoring signals to be processed.
The step 1 comprises the following steps:
step 1.1: collecting a monitoring signal when no blasting occurs in a mine site, screening out a micro-seismic signal by comparing the monitoring signal with an ideal waveform of the micro-seismic signal, and establishing a subdata set of the micro-seismic signal;
step 1.2: collecting a monitoring signal when mine field blasting occurs, comparing the monitoring signal with an ideal waveform of a blasting signal, screening the blasting signal, and establishing a subdata set of the blasting signal;
step 1.3: collecting monitoring signals in an operation area of mechanical equipment in a mine field, screening out mechanical vibration signals through comparison with ideal waveforms of the mechanical vibration signals, and establishing a subdata set of the mechanical vibration signals;
step 1.4: and constructing a waveform database according to the sub data sets of the microseismic signal, the blasting signal and the mechanical vibration signal.
The method is characterized in that each RGB picture is endowed with a label, and the specific expression is as follows: and assigning the RGB pictures corresponding to all the microseismic signals as microseismic labels, assigning the RGB pictures corresponding to all the blasting signals as blasting labels, and assigning the RGB pictures corresponding to all the mechanical vibration signals as mechanical vibration labels.
The invention has the beneficial effects that:
the invention provides a mine micro-seismic signal automatic identification method based on deep learning, which combines signal waveform time-frequency characteristics and a deep learning method for the first time and is applied to mine micro-seismic signal identification. By using a short-time Fourier transform (STFT) method, the method does not need to extract parameters like a conventional method, and can take essential parameters, time, frequency and amplitude characteristics of the waveform into consideration by using one-time STFT. And selecting the signal time-frequency characteristic diagram after STFT conversion as the input of a convolutional neural network (CNN for short), and obtaining a high-quality microseismic signal identification effect with higher accuracy by training the CNN.
The mine microseismic signal is identified by adopting the method and four classification methods of a common support vector machine, a decision tree, a K-means neighbor algorithm and linear discriminant analysis, and the result shows that the identification accuracy rate of the method is far higher than that of the other four methods, and the microseismic signal can be accurately identified from the monitoring signal of the complex mine underground environment.
Drawings
Fig. 1 is a flow chart of an automatic mine microseismic signal identification method based on deep learning in the invention.
FIG. 2 is a time-amplitude diagram of microseismic signals, (b) a frequency-amplitude diagram of the microseismic signals after STFT, (c) a time-amplitude diagram of blasting signals, (d) a frequency-amplitude diagram of the blasting signals after STFT, (e) a time-amplitude diagram of mechanical vibration signals, and (f) a frequency-amplitude diagram of the mechanical vibration signals after STFT;
FIG. 3 is a topological structure diagram of a convolutional neural network designed by the present invention;
FIG. 4 is a schematic diagram of a fully-connected layer and a softmax regression layer in a convolutional neural network designed by the present invention;
FIG. 5 is a graph of the classification results using the method of the present invention, wherein (a) is a graph of the accuracy change of the training process, and (b) is an error matrix graph;
FIG. 6 is a chart comparing the accuracy of each classification method, wherein: MS is a microseismic signal; b is a blasting signal; m is a mechanical vibration signal; SVM is a support vector machine method; DT is a decision tree method; KNN is a K-means neighbor method; LD is a linear discriminant analysis method.
Detailed Description
The invention is further described with reference to the following figures and specific examples.
In order to improve the identification rate and accuracy of mine microseismic signals, the invention firstly acquires monitoring signals through a mine underground monitoring system, because the directly acquired monitoring signals not only comprise required microseismic signals, but also comprise blasting signals when field blasting occurs and mechanical vibration signals generated when field mechanical equipment works, three signals need to be identified when a sample set is constructed, then the waveforms of the three signals are subjected to short-time Fourier transform (STFT) to obtain a time-frequency characteristic diagram, and the time-frequency characteristic diagram is organically combined with a convolutional neural network (CNN for short) in a deep learning method, so that the mine microseismic signal automatic identification method based on deep learning is provided, as shown in figure 1, the method comprises the following steps:
step 1: establishing a waveform database containing a microseismic signal, a blasting signal and a mechanical vibration signal by a manual method, comparing the recorded blasting with the microseismic signal according to the blasting time and the blasting place recorded in a mine site, manually screening the blasting waveform, and establishing a blasting signal database; screening signals obtained in a time period when the mine has no blasting and mechanical vibration signals are few, and establishing a microseismic signal database; the method comprises the following steps of manually judging the waveform to obtain a mechanical vibration signal, and establishing a mechanical vibration signal database, wherein the method specifically comprises the following steps:
step 1.1: collecting a monitoring signal when no blasting occurs in a mine site, screening out a micro-seismic signal by comparing the monitoring signal with an ideal waveform of the micro-seismic signal, and establishing a subdata set of the micro-seismic signal;
step 1.2: collecting a monitoring signal when mine field blasting occurs, comparing the monitoring signal with an ideal waveform of a blasting signal, screening the blasting signal, and establishing a subdata set of the blasting signal;
step 1.3: collecting monitoring signals in an operation area of mechanical equipment in a mine field, screening out mechanical vibration signals through comparison with ideal waveforms of the mechanical vibration signals, and establishing a subdata set of the mechanical vibration signals;
step 1.4: constructing a waveform database according to the subdata sets of the microseismic signal, the blasting signal and the mechanical vibration signal;
the waveform database created by manual methods contains 2000 microseismic waveforms, 2000 blasting waveforms and 2000 mechanical vibration waveforms.
Step 2: processing all waveform data (namely microseismic signals, blasting signals and mechanical vibration signals) in a waveform database into the same waveform length by adopting a down-sampling method to obtain a preprocessed signal waveform;
and step 3: generating two-dimensional RGB pictures by adopting short-time Fourier transform on the preprocessed signal waveforms, and taking all the RGB pictures as a sample set; fig. 2 is a time-frequency-amplitude diagram before and after the microseismic signal, the explosion signal and the mechanical vibration signal are subjected to short-time fourier transform, the RGB picture structure is 227 × 227 × 3, that is, the resolution of the picture is 227 × 227, and each pixel point corresponds to a vector of RGB three-component colors.
And 4, step 4: the sample set is divided into a training set and a verification set according to a preset proportion, and data of 70% of various waveforms are randomly selected as training samples, namely 1400 waveforms of the microseismic signal, the blasting signal and the mechanical vibration signal.
And 5: determining the number of the convolutional layers, the fully-connected layers and the number of internal neurons by using a trial-and-error method, and constructing a convolutional neural network model;
in order to determine the optimal number of convolutional layers and corresponding output feature maps in the CNN, a plurality of tests of different structures are performed, and finally, it is determined that the CNN topology is as shown in fig. 3. The specific design parameters are as follows: the convolution Kernel (Kernel) size of the first convolutional layer C1 is 11 × 11 × 3, the number is 96, the sliding step is 4, 96 feature maps with the size of 55 × 55 are obtained after convolution operation, then downsampling is performed on the first pooling layer, the size of a pooling Window (Window) is 3 × 3, the Window sliding step is 2, the maximum pooling operation is adopted, and the feature map size is reduced to 27 × 27; then, the method enters a second convolution layer C2, the size of convolution kernels is 5 multiplied by 5, the number of the convolution kernels is 256, the sliding step length is 1, and 256 characteristic graphs of 27 multiplied by 27 are obtained after convolution operation; entering a second pooling layer M2, wherein pooling parameters are the same as those of the pooling layer M1, and the size of the feature map is reduced to 13 multiplied by 13; then, through three layers of convolutional layers (C3, C4 and C5), the sizes of convolutional kernels are all 3 × 3, the sliding step lengths are all 1, and finally 256 4 × 4 highly abstract feature maps are obtained; finally, features of three full-connection layers (F6, F7 and F8) are combined to obtain a 1000 × 1 vector, and the vector is transferred to a Softmax regression layer to be classified, as shown in FIG. 4. And when the classification result is inconsistent with the given label, adopting a back propagation training model to finally obtain the CNN model for identifying the microseismic signal.
In the CNN topology shown in fig. 3, C1, C2, C3, C4 and C5 are respectively a first convolutional layer, a second convolutional layer, a third convolutional layer, a fourth convolutional layer and a fifth convolutional layer; m1, M2, M3, M4 and M5 are respectively a first pooling layer, a second pooling layer, a third pooling layer, a fourth pooling layer and a fifth pooling layer; f6, F7, F8 are three fully connected layers, the pooling layer and the Softmax regression layer are shown in fig. 4.
Step 6: taking the RGB pictures corresponding to each waveform in the training set as input, and assigning a label to each RGB picture when inputting, namely assigning the RGB pictures corresponding to all microseismic signals as microseismic labels, assigning the RGB pictures corresponding to all blasting signals as blasting labels, and assigning the RGB pictures corresponding to all mechanical vibration signals as mechanical vibration labels;
and 7: training the convolutional neural network model through a training set to obtain a trained model;
and 8: taking the RGB pictures corresponding to each waveform in the verification set as input, assigning a label to each RGB picture when inputting, namely assigning the RGB pictures corresponding to all microseismic signals as microseismic labels, assigning the RGB pictures corresponding to all blasting signals as blasting labels, assigning the RGB pictures corresponding to all mechanical vibration signals as mechanical vibration labels, and verifying the trained model through the verification set to obtain the accuracy of the model in signal recognition;
and step 9: the accuracy is more than or equal to a set threshold value epsilon2The model is used as a parameter optimal model, and the parameter optimal model is used for automatically identifying the microseismic signals of the monitoring signals to be processed.
Fig. 5(a) is a curve of the accuracy of signal identification by the method of the present invention and the variation of the loss function with the number of iteration steps, and it can be seen that, when the number of iteration steps is 100, the average accuracy of identification of three waveforms is already over 90%. When the steps are iterated to 1050, the accuracy and the loss function tend to be stable, training is terminated, and finally the average recognition accuracy of the three waveforms reaches 94.73%; as can be seen from the error matrix of the STFT-CNN given in FIG. 5(b), the discrimination rate of the blasting and microseismic signals is excellent and reaches more than 97%. Comparing the signal recognition result of the invention with the results obtained by the signal recognition and classification methods of the common support vector machine (SVM for short), the decision tree (DT for short), the K-means neighbor algorithm (KNN for short) and the linear discriminant analysis (LD for short) respectively, as shown in FIG. 6, the overall accuracy of the method of the invention is improved by 9.76 percent compared with the SVM method with the best recognition effect in the common method; in addition, when the mechanical vibration and the microseismic signals are distinguished by the common method, due to the fact that the parameters of the two are similar, the classification effect is poor, the mechanical vibration signal recognition rate of the SVM method is only 70.8%, the DT method is 71.7%, the KNN method is 73.8% and the LD method is 70.7%, the recognition rate of the mechanical vibration signal by adopting the method (STFT-CNN for short) of the invention is 89.6%, and only 0.5% of the mechanical vibration signal can be mistaken for the microseismic signal, and the recognition rate is obviously improved.
Claims (3)
1. A mine microseismic signal automatic identification method based on deep learning is characterized by comprising the following steps:
step 1: establishing a waveform database containing microseismic signals, blasting signals and mechanical vibration signals;
step 2: processing all waveform data in a waveform database into the same waveform length by adopting a downsampling method to obtain a preprocessed signal waveform;
and step 3: generating two-dimensional RGB pictures by adopting short-time Fourier transform on the preprocessed signal waveforms, and taking all the RGB pictures as a sample set;
and 4, step 4: dividing a sample set into a training set and a verification set according to a preset proportion;
and 5: determining the number of the convolutional layers, the fully-connected layers and the number of internal neurons by using a trial-and-error method, and constructing a convolutional neural network model;
step 6: taking the RGB picture corresponding to each waveform in the training set as input, and endowing each RGB picture with a label during input;
and 7: training the convolutional neural network model through a training set to obtain a trained model;
and 8: taking the RGB pictures corresponding to each waveform in the verification set as input, endowing each RGB picture with a label when inputting, and verifying the trained model through the verification set to obtain the accuracy rate of the model for signal recognition;
and step 9: the accuracy is more than or equal to a set threshold value epsilon2The model is used as a parameter optimal model, and the parameter optimal model is used for automatically identifying the microseismic signals of the monitoring signals to be processed.
2. The method for automatically identifying mine microseismic signals based on deep learning of claim 1 wherein the step 1 comprises:
step 1.1: collecting a monitoring signal when no blasting occurs in a mine site, screening out a micro-seismic signal by comparing the monitoring signal with an ideal waveform of the micro-seismic signal, and establishing a subdata set of the micro-seismic signal;
step 1.2: collecting a monitoring signal when mine field blasting occurs, comparing the monitoring signal with an ideal waveform of a blasting signal, screening the blasting signal, and establishing a subdata set of the blasting signal;
step 1.3: collecting monitoring signals in an operation area of mechanical equipment in a mine field, screening out mechanical vibration signals through comparison with ideal waveforms of the mechanical vibration signals, and establishing a subdata set of the mechanical vibration signals;
step 1.4: and constructing a waveform database according to the sub data sets of the microseismic signal, the blasting signal and the mechanical vibration signal.
3. The mine microseismic signal automatic identification method based on deep learning of claim 1 wherein each RGB picture is assigned a label specifically expressed as: and assigning the RGB pictures corresponding to all the microseismic signals as microseismic labels, assigning the RGB pictures corresponding to all the blasting signals as blasting labels, and assigning the RGB pictures corresponding to all the mechanical vibration signals as mechanical vibration labels.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011357051.1A CN112487952A (en) | 2020-11-27 | 2020-11-27 | Mine microseismic signal automatic identification method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011357051.1A CN112487952A (en) | 2020-11-27 | 2020-11-27 | Mine microseismic signal automatic identification method based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112487952A true CN112487952A (en) | 2021-03-12 |
Family
ID=74936133
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011357051.1A Pending CN112487952A (en) | 2020-11-27 | 2020-11-27 | Mine microseismic signal automatic identification method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112487952A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117611894A (en) * | 2023-11-24 | 2024-02-27 | 中国矿业大学 | Microseism effective signal intelligent recognition method and system based on self-supervision learning |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108846307A (en) * | 2018-04-12 | 2018-11-20 | 中南大学 | A kind of microseism based on waveform image and explosion events recognition methods |
| CN110058294A (en) * | 2019-05-10 | 2019-07-26 | 东北大学 | A kind of tunnel micro seismic monitoring rock rupture event automatic identifying method |
| CN110632662A (en) * | 2019-09-25 | 2019-12-31 | 成都理工大学 | Algorithm for automatically identifying microseism signals by using DCNN-inclusion network |
| CN110866448A (en) * | 2019-10-21 | 2020-03-06 | 西北工业大学 | Flutter signal analysis method based on convolutional neural network and short-time Fourier transform |
| CN111562612A (en) * | 2020-05-20 | 2020-08-21 | 大连理工大学 | Deep learning microseismic event identification method and system based on attention mechanism |
-
2020
- 2020-11-27 CN CN202011357051.1A patent/CN112487952A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108846307A (en) * | 2018-04-12 | 2018-11-20 | 中南大学 | A kind of microseism based on waveform image and explosion events recognition methods |
| CN110058294A (en) * | 2019-05-10 | 2019-07-26 | 东北大学 | A kind of tunnel micro seismic monitoring rock rupture event automatic identifying method |
| CN110632662A (en) * | 2019-09-25 | 2019-12-31 | 成都理工大学 | Algorithm for automatically identifying microseism signals by using DCNN-inclusion network |
| CN110866448A (en) * | 2019-10-21 | 2020-03-06 | 西北工业大学 | Flutter signal analysis method based on convolutional neural network and short-time Fourier transform |
| CN111562612A (en) * | 2020-05-20 | 2020-08-21 | 大连理工大学 | Deep learning microseismic event identification method and system based on attention mechanism |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117611894A (en) * | 2023-11-24 | 2024-02-27 | 中国矿业大学 | Microseism effective signal intelligent recognition method and system based on self-supervision learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Misra et al. | Noninvasive fracture characterization based on the classification of sonic wave travel times | |
| US9519865B2 (en) | Apparatus and methods of analysis of pipe and annulus in a wellbore | |
| Del Pezzo et al. | Discrimination of earthquakes and underwater explosions using neural networks | |
| Esposito et al. | Automatic discrimination among landslide, explosion-quake, and microtremor seismic signals at Stromboli volcano using neural networks | |
| GB2590555A (en) | Methods for characterizing and evaluating well integrity using unsupervised machine learning of acoustic data | |
| CN111126332B (en) | Frequency hopping signal classification method based on contour features | |
| CN110440148B (en) | Method, device and system for classifying and identifying leakage acoustic signals | |
| CN110632662A (en) | Algorithm for automatically identifying microseism signals by using DCNN-inclusion network | |
| CN110568483A (en) | Automatic evaluation method for seismic linear noise suppression effect based on convolutional neural network | |
| Lara-Cueva et al. | On the use of multi-class support vector machines for classification of seismic signals at Cotopaxi volcano | |
| CN113792685A (en) | Microseism event detection method based on multi-scale convolution neural network | |
| CN115700542A (en) | Optical fiber pipeline safety early warning algorithm based on deep learning | |
| CN112487952A (en) | Mine microseismic signal automatic identification method based on deep learning | |
| CN111894563B (en) | Classification determination method and system for crack type reservoir in submarine mountain section | |
| CN117292148B (en) | Tunnel surrounding rock level assessment method based on directional drilling and test data | |
| CN116299684B (en) | Novel microseismic classification method based on bimodal neurons in artificial neural network | |
| CN110688983A (en) | Microseismic signal identification method based on multi-mode optimization and ensemble learning | |
| CN111983569B (en) | Radar interference suppression method based on neural network | |
| CN115563480A (en) | Gear fault identification method for screening octave geometric modal decomposition based on kurtosis ratio coefficient | |
| CN112412390B (en) | Method and device for evaluating second interface of well cementation based on deep learning model | |
| CN113901878A (en) | CNN + RNN algorithm-based three-dimensional ground penetrating radar image underground pipeline identification method | |
| KR102418118B1 (en) | Apparatus and method of deep learning-based facility diagnosis using frequency synthesis | |
| CN116628549A (en) | Electromagnetic method signal-to-noise identification method and system based on extreme learning machine | |
| CN114624770A (en) | Seismic surface wave detection method based on convolutional neural network | |
| CN112699788A (en) | Microseism P wave polarity identification method, device, storage medium and equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |