CN115220094A - Microseism event detection method based on BiLSTM and attention mechanism - Google Patents
Microseism event detection method based on BiLSTM and attention mechanism Download PDFInfo
- Publication number
- CN115220094A CN115220094A CN202210842326.3A CN202210842326A CN115220094A CN 115220094 A CN115220094 A CN 115220094A CN 202210842326 A CN202210842326 A CN 202210842326A CN 115220094 A CN115220094 A CN 115220094A
- Authority
- CN
- China
- Prior art keywords
- bilstm
- network
- features
- attention mechanism
- layer
- 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.)
- Granted
Links
- 230000007246 mechanism Effects 0.000 title claims abstract description 47
- 238000001514 detection method Methods 0.000 title claims abstract description 20
- 238000000034 method Methods 0.000 claims abstract description 33
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 23
- 238000012545 processing Methods 0.000 claims abstract description 12
- 230000004927 fusion Effects 0.000 claims abstract description 9
- 238000012544 monitoring process Methods 0.000 claims abstract description 7
- 238000010606 normalization Methods 0.000 claims description 12
- 230000004913 activation Effects 0.000 claims description 8
- 238000013528 artificial neural network Methods 0.000 claims description 7
- 239000000284 extract Substances 0.000 claims description 7
- 230000009466 transformation Effects 0.000 claims description 7
- 238000010276 construction Methods 0.000 claims description 4
- 238000010586 diagram Methods 0.000 description 8
- 238000000605 extraction Methods 0.000 description 4
- 238000013135 deep learning Methods 0.000 description 2
- 238000003058 natural language processing Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V1/00—Seismology; Seismic or acoustic prospecting or detecting
- G01V1/28—Processing seismic data, e.g. for interpretation or for event detection
- G01V1/288—Event detection in seismic signals, e.g. microseismics
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Environmental & Geological Engineering (AREA)
- Remote Sensing (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- General Life Sciences & Earth Sciences (AREA)
- Acoustics & Sound (AREA)
- Geophysics (AREA)
- Business, Economics & Management (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Geology (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Emergency Management (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a microseism event detection method based on a BilSTM and attention mechanism, which comprises the following steps: acquiring time sequence data acquired by a microseism monitoring instrument; converting the time sequence data into frequency domain data, storing the frequency domain data in a spectrogram form, and extracting shallow features of the spectrogram based on a convolutional neural network; inputting the shallow feature into a BilSt network to obtain a first deep feature, and inputting the shallow feature into a multi-attention mechanism network to obtain a second deep feature; and performing feature fusion on the first deep features and the second deep features, and inputting the feature fusion into the multilayer perceptron network to obtain a detection result. The method can improve the attention to the overall attributes of the microseism data, improve the capability of processing the microseism data and effectively judge the data containing effective microseism events in the frequency domain.
Description
Technical Field
The invention belongs to the field of microseism event detection, and particularly relates to a microseism event detection method based on a BilSTM and attention mechanism.
Background
The hydraulic fracturing technology is widely applied to monitoring of oil field production increase and exploitation of new energy such as shale gas. By deploying microseismic monitoring equipment in the well and at the surface, microseismic events due to hydraulic fracturing can be recorded. As the size of microseismic monitoring increases, the amount of data collected on site increases, and data processing and analysis requires that as much useful information as possible be extracted from the data. The deep learning technique is an effective method widely used and is suitable for processing a large amount of data, which makes it suitable for use in the processing of micro-seismic data. The existing deep learning-based research is mostly dedicated to solving the related tasks of processing the micro-seismic data, and similar features in the micro-seismic data and the correlation among the features are not well utilized.
The multi-head attention mechanism is applied to natural language processing for the earliest time and is rapidly developed. The multi-head attention mechanism can better analyze the global dependency of input data. In the fields of natural language processing and images, there have been many studies and in some tasks the performance of recurrent and convolutional neural networks has been exceeded. Seismic data may be considered similar to audio (time series) data or image data in that it contains some attributes of both.
Disclosure of Invention
The invention aims to provide a microseismic event detection method based on BilSTM and attention mechanism, which solves the problems in the prior art.
In order to achieve the purpose, the invention provides a microseism event detection method based on BilSTM and attention mechanism, which comprises the following steps:
acquiring time sequence data acquired by a microseism monitoring instrument; converting the time sequence data into frequency domain data, storing the frequency domain data in a spectrogram form, constructing a convolutional neural network, and extracting shallow features of the spectrogram based on the convolutional neural network; constructing a BilSTM network, inputting the shallow feature into the BilSTM network to obtain a first deep feature, and inputting the shallow feature into a multi-attention mechanism network to obtain a second deep feature; and performing feature fusion on the first deep-layer features and the second deep-layer features, and inputting the fused features into a multi-layer perceptron network to obtain a detection result.
Optionally, the convolutional neural network includes: a plurality of two-dimensional convolutional layers, a batch normalization layer, and an activation layer.
Optionally, the plurality of two-dimensional convolutional layers comprises: the convolution kernel 1 is a two-dimensional convolution with 9 × 9, the convolution kernel 3 is a two-dimensional convolution with 7 × 7, the convolution kernel 1 is a two-dimensional convolution with 5 × 5, and the convolution kernel 2 is a two-dimensional convolution with 3 × 3, wherein different convolution kernels are connected in the form of residual errors.
Optionally, the constructing of the convolutional neural network includes: the two-dimensional convolutions with the convolution kernels of 7 × 7 are respectively connected with batch standardized layers and are stacked in sequence, the two-dimensional convolution with the convolution kernels of 9 × 9 is connected with batch standardized layers connected with the two-dimensional convolution with the convolution kernels of 7 × 7, the two-dimensional convolution with the convolution kernels of 1 × 5 is connected with the two-dimensional convolutions with the convolution kernels of 3 × 3, the two-dimensional convolutions with the convolution kernels of 2 × 3 are respectively connected with the batch standardized layers and are stacked in sequence, the two-dimensional convolution with the convolution kernels of 5 × 5 is connected with the two-dimensional convolution with the convolution kernels of 7 × 7, and the batch standardized layers connected with the two-dimensional convolution with the convolution kernels of 3 × 3, and any batch of standardized layers are located above the two-dimensional convolutions.
Optionally, the obtaining of the first deep feature includes: and redistributing the shallow layer features into a sequence, inputting the sequence into a BilSTM network, and extracting features from two directions by the BilSTM network, wherein the BilSTM network comprises two BilSTM layers, three batch normalization layers and a one-way LSTM layer.
Optionally, the constructing of the BiLSTM network includes: the two BiLSTM layers are respectively connected with the batch standardization layer, and the one-way LSTM layer is connected with the batch standardization layer connected with the second BiLSTM layer and the third batch standardization layer.
Optionally, the obtaining of the second deep feature includes: and performing linear transformation on the shallow features, inputting the shallow features into a multi-head attention mechanism network, and acquiring the second deep features, wherein the multi-head attention mechanism network comprises N multi-head attention mechanism network layers and a feedforward neural network.
Optionally, the operation process of the multi-head attention mechanism network includes: and summing and standardizing the linearly transformed shallow features after the multi-head attention mechanism network layer processing and the linearly transformed shallow features to obtain process features, and summing and standardizing the process features and the process features after the feedforward neural network processing to obtain the second deep features.
Optionally, the multi-layer perceptron network includes a plurality of fully-connected layers and an active layer.
The invention has the technical effects that:
the method extracts shallow features of the micro-seismic data by applying a multilayer convolutional neural network, extracts deep features of the micro-seismic data by applying a BilSTM network and a multi-head attention mechanism network in parallel, and performs fusion judgment on the depth features extracted by the two networks in parallel. The method can improve the attention to the overall attributes of the microseism data, improve the capability of processing the microseism data and effectively judge the data containing effective microseism events in the frequency domain.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate embodiments of the application and, together with the description, serve to explain the application and are not intended to limit the application. In the drawings:
FIG. 1 is a block diagram of a method for microseismic event detection based on BilSTM and attention mechanism in an embodiment of the present invention;
FIG. 2 is a block diagram of shallow feature extraction based on convolutional neural network in an embodiment of the present invention;
FIG. 3 is a block diagram of deep feature extraction based on a BilSTM network in an embodiment of the present invention;
FIG. 4 is a block diagram of deep feature extraction based on a multi-head attention mechanism in an embodiment of the present invention;
fig. 5 is a block diagram of parallel network feature fusion and multi-layer perceptron detection in an embodiment of the present invention.
Detailed Description
It should be noted that, in the present application, the embodiments and features of the embodiments may be combined with each other without conflict. The present application will be described in detail below with reference to the embodiments with reference to the attached drawings.
It should be noted that the steps illustrated in the flowcharts of the figures may be performed in a computer system such as a set of computer-executable instructions and that, although a logical order is illustrated in the flowcharts, in some cases, the steps illustrated or described may be performed in an order different than presented herein.
Example one
As shown in fig. 1, the present embodiment provides a method for detecting a micro-seismic event based on BiLSTM and attention mechanism, which includes:
frequency domain transformation, a convolutional neural network layer, a BilSTM network layer, a multi-head attention mechanism network and a multi-layer perceptron network.
The frequency domain transformation is to transform the time domain data to the frequency domain by applying a frequency domain changing method such as fast Fourier transformation, short-time Fourier transformation and the like. The transformed frequency domain data is stored in the form of a spectrogram.
And the convolutional neural network layer is used for extracting shallow layer characteristics of the input spectrogram.
And the BilSTM network layer and the multi-head attention mechanism network perform further feature extraction on the output of the convolutional neural network in parallel. The features extracted by the two layers are fused and then sent to a multilayer perceptron network.
And the multilayer perceptron network classifies and judges the fused features.
The convolutional neural network layer is configured as a block diagram, as shown in fig. 2, and includes a plurality of convolutional layers, batch normalization layers, and activation layers.
The convolutional neural network is composed of a plurality of two-dimensional convolutional layers, a batch normalization layer and an activation layer. Convolution kernels of different sizes are connected in a residual situation. The application of the batch normalization layer helps to prevent overfitting of the network and can speed up the training process.
The convolutional neural network layer is composed of 1 two-dimensional convolution with convolution kernels of 9 × 9, 3 two-dimensional convolutions with convolution kernels of 7 × 7, 1 two-dimensional convolution with convolution kernels of 5 × 5, 2 two-dimensional convolutions with convolution kernels of 3 × 3, and corresponding batch of normalization layers and activation layers. Through training, the network layers can automatically extract shallow features in the spectrogram.
The block diagram of the BilSTM network layer is shown in FIG. 3, and comprises a BilSTM layer, a batch normalization layer and an LSTM layer.
In a BiLSTM network, shallow features extracted by the convolutional neural network layer are reassigned into sequences.
The sequence was passed into a two-layer network consisting of BilSTM and batch normalization layers.
BilSTM extracts features from two directions, and can extract and represent features related to each other in time series.
Since the seismic data is recorded unidirectionally, an additional unidirectional LSTM layer is added after the two BiLSTM layers to map the learned features of the first two bi-directional layers.
The multi-head attention mechanism network is a block diagram, as shown in fig. 4, composed of N multi-head attention mechanism networks and a feedforward neural network stack.
In the multi-head attention mechanism network, the network can perform global operation, and one multi-head attention mechanism network layer can model the relation among all pixels, namely can better extract hidden features in data.
And the deep features extracted by the BilSTM network and the deep features extracted by the multi-head attention mechanism network are subjected to feature fusion and then are sent to the multilayer perceptron network.
The multi-layer perceptron network is constructed as shown in fig. 5. The system consists of a plurality of full connection layers and activation layers, and can classify and judge the fused features.
And finally, outputting a detection result, namely, which data contain valid microseismic events and which data are noise data.
Example two
The embodiment provides a microseism event detection method based on BilSTM and attention mechanism, which comprises the following steps:
acquiring time sequence data acquired by a microseism monitoring instrument; converting the time sequence data into frequency domain data, storing the frequency domain data in a spectrogram form, constructing a convolutional neural network, and extracting shallow features of the spectrogram based on the convolutional neural network; constructing a BilSTM network, inputting the shallow layer characteristics into the BilSTM network to obtain first deep layer characteristics, and inputting the shallow layer characteristics into a multi-attention mechanism network to obtain second deep layer characteristics; and performing feature fusion on the first deep features and the second deep features, and inputting the feature fusion into the multilayer perceptron network to obtain a detection result.
In some embodiments, the convolutional neural network comprises: a plurality of two-dimensional convolutional layers, a batch normalization layer, and an activation layer.
In some embodiments, the plurality of two-dimensional convolutional layers comprises: the convolution kernel 1 is a two-dimensional convolution with 9 × 9, the convolution kernel 3 is a two-dimensional convolution with 7 × 7, the convolution kernel 1 is a two-dimensional convolution with 5 × 5, and the convolution kernel 2 is a two-dimensional convolution with 3 × 3, wherein different convolution kernels are connected in the form of residual errors.
In some embodiments, the construction of the convolutional neural network comprises: the two-dimensional convolutions with the convolution kernels of 7 × 7 are respectively connected with batch standardized layers and are stacked in sequence, the two-dimensional convolution with the convolution kernels of 9 × 9 is connected with batch standardized layers connected with the two-dimensional convolution with the convolution kernels of 7 × 7, the two-dimensional convolution with the convolution kernels of 1 × 5 is connected with the two-dimensional convolutions with the convolution kernels of 3 × 3, the two-dimensional convolutions with the convolution kernels of 2 × 3 are respectively connected with the batch standardized layers and are stacked in sequence, the two-dimensional convolution with the convolution kernels of 5 × 5 is connected with the two-dimensional convolution with the convolution kernels of 7 × 7, and the batch standardized layers connected with the two-dimensional convolution with the convolution kernels of 3 × 3, and any batch of standardized layers are located above the two-dimensional convolutions.
In some embodiments, the process of obtaining the first deep feature comprises: and redistributing the shallow layer features into a sequence, inputting the sequence into a BilSTM network, and extracting the features from two directions by the BilSTM network, wherein the BilSTM network comprises two BilSTM layers, three batch normalization layers and a one-way LSTM layer.
In some embodiments, the construction of the BilStm network comprises: the two BiLSTM layers are respectively connected with the batch standardization layer, and the one-way LSTM layer is connected with the batch standardization layer connected with the second BiLSTM layer and the third batch standardization layer.
In some embodiments, the process of obtaining the second deep features includes: and performing linear transformation on the shallow features, and inputting the shallow features into a multi-head attention mechanism network to obtain a second deep feature, wherein the multi-head attention mechanism network comprises N multi-head attention mechanism network layers and a feedforward neural network.
In some embodiments, the multi-head attention mechanism network operation process comprises: and summing and standardizing the linearly transformed shallow features after the multi-head attention mechanism network layer processing and the linearly transformed shallow features to obtain process features, and summing and standardizing the process features and the process features after the feedforward neural network processing to obtain a second deep feature.
In some embodiments, the multi-tier perceptron network includes several fully-connected tiers and an activation tier.
The above description is only for the preferred embodiment of the present application, but the scope of the present application is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present application should be covered within the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (9)
1. A microseism event detection method based on a BilSTM and attention mechanism is characterized by comprising the following steps:
acquiring time sequence data acquired by a microseism monitoring instrument; converting the time sequence data into frequency domain data, storing the frequency domain data in a spectrogram form, constructing a convolutional neural network, and extracting shallow features of the spectrogram based on the convolutional neural network; constructing a BilSTM network, inputting the shallow feature into the BilSTM network to obtain a first deep feature, and inputting the shallow feature into a multi-attention mechanism network to obtain a second deep feature; and performing feature fusion on the first deep-layer features and the second deep-layer features, and inputting the fused features into a multi-layer perceptron network to obtain a detection result.
2. The method for microseismic event detection based on BilSTM and attention mechanism of claim 1 wherein the convolutional neural network comprises: a plurality of two-dimensional convolutional layers, a batch normalization layer, and an activation layer.
3. The method of claim 2, wherein the plurality of two-dimensional convolutional layers comprise: the convolution kernel 1 is a two-dimensional convolution with 9 × 9, the convolution kernel 3 is a two-dimensional convolution with 7 × 7, the convolution kernel 1 is a two-dimensional convolution with 5 × 5, and the convolution kernel 2 is a two-dimensional convolution with 3 × 3, wherein different convolution kernels are connected in the form of residual errors.
4. The method for detecting microseismic events based on BilSTM and attention mechanism of claim 3 wherein the construction of the convolutional neural network comprises: the two-dimensional convolutions with convolution kernels of 7 × 7 of 3 are respectively connected to the batch of normalized layers and stacked in sequence, the two-dimensional convolution with convolution kernels of 9 × 9 is connected to the batch of normalized layers connected to the two-dimensional convolution with convolution kernels of 7 × 7, and the two-dimensional convolutions with convolution kernels of 5 × 5 of 1, the two-dimensional convolutions with convolution kernels of 3 × 3 of 2 are respectively connected to the batch of normalized layers and stacked in sequence, the two-dimensional convolution with convolution kernels of 5 × 5 is connected to the two-dimensional convolution with convolution kernels of 7 of the third and the batch of normalized layers connected to the two-dimensional convolution with convolution kernels of 3 × 3, wherein any batch of normalized layers is located above the two-dimensional convolution.
5. The method of claim 1 for microseismic event detection based on BilSTM and attention mechanism, wherein the process of obtaining the first deep signature comprises: and redistributing the shallow layer features into a sequence, and inputting the sequence into a BilSTM network, wherein the BilSTM network extracts features from two directions, and the BilSTM network comprises two BilSTM layers, three batch normalization layers and a one-way LSTM layer.
6. The method of claim 5 for microseismic event detection based on BilSTM and attention mechanisms, wherein the construction of the BilSTM network comprises: the two BiLSTM layers are respectively connected with the batch standardization layer, and the one-way LSTM layer is connected with the batch standardization layer connected with the second BiLSTM layer and the third batch standardization layer.
7. The method of claim 1 for microseismic event detection based on BilSTM and attention mechanism, wherein the process of obtaining the second deep layer features comprises: and performing linear transformation on the shallow features, and inputting the shallow features into a multi-head attention mechanism network to obtain the second deep features, wherein the multi-head attention mechanism network comprises N multi-head attention mechanism network layers and a feedforward neural network.
8. The method for detecting microseismic events based on BilSTM and attention mechanism as claimed in claim 1 wherein the multi-head attention mechanism network operation process comprises: and summing and standardizing the linearly transformed shallow features after the multi-head attention mechanism network layer processing and the linearly transformed shallow features to obtain process features, and summing and standardizing the process features and the process features after the feedforward neural network processing to obtain the second deep features.
9. The method for detecting microseismic events based on BilSTM and attention mechanism of claim 1 wherein the multi-layered perceptron network comprises a plurality of fully connected and active layers.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210842326.3A CN115220094B (en) | 2022-07-18 | 2022-07-18 | Microseism event detection method based on BilSTM and attention mechanism |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210842326.3A CN115220094B (en) | 2022-07-18 | 2022-07-18 | Microseism event detection method based on BilSTM and attention mechanism |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115220094A true CN115220094A (en) | 2022-10-21 |
| CN115220094B CN115220094B (en) | 2023-01-17 |
Family
ID=83611489
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210842326.3A Active CN115220094B (en) | 2022-07-18 | 2022-07-18 | Microseism event detection method based on BilSTM and attention mechanism |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115220094B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116660992A (en) * | 2023-06-05 | 2023-08-29 | 北京石油化工学院 | Seismic signal processing method based on multi-feature fusion |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111651504A (en) * | 2020-06-03 | 2020-09-11 | 湖南大学 | Multi-element time sequence multilayer space-time dependence modeling method based on deep learning |
| CN112464721A (en) * | 2020-10-28 | 2021-03-09 | 中国石油天然气集团有限公司 | Automatic microseism event identification method and device |
| US20210117737A1 (en) * | 2019-10-18 | 2021-04-22 | Korea University Research And Business Foundation | Earthquake event classification method using attention-based convolutional neural network, recording medium and device for performing the method |
| CN113985475A (en) * | 2021-10-28 | 2022-01-28 | 北京石油化工学院 | Micro-seismic monitoring data transmission method based on terminal edge cloud cooperation of Internet of things |
| WO2022121150A1 (en) * | 2020-12-10 | 2022-06-16 | 平安科技（深圳）有限公司 | Speech recognition method and apparatus based on self-attention mechanism and memory network |
-
2022
- 2022-07-18 CN CN202210842326.3A patent/CN115220094B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20210117737A1 (en) * | 2019-10-18 | 2021-04-22 | Korea University Research And Business Foundation | Earthquake event classification method using attention-based convolutional neural network, recording medium and device for performing the method |
| CN111651504A (en) * | 2020-06-03 | 2020-09-11 | 湖南大学 | Multi-element time sequence multilayer space-time dependence modeling method based on deep learning |
| CN112464721A (en) * | 2020-10-28 | 2021-03-09 | 中国石油天然气集团有限公司 | Automatic microseism event identification method and device |
| WO2022121150A1 (en) * | 2020-12-10 | 2022-06-16 | 平安科技（深圳）有限公司 | Speech recognition method and apparatus based on self-attention mechanism and memory network |
| CN113985475A (en) * | 2021-10-28 | 2022-01-28 | 北京石油化工学院 | Micro-seismic monitoring data transmission method based on terminal edge cloud cooperation of Internet of things |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116660992A (en) * | 2023-06-05 | 2023-08-29 | 北京石油化工学院 | Seismic signal processing method based on multi-feature fusion |
| CN116660992B (en) * | 2023-06-05 | 2024-03-05 | 北京石油化工学院 | Seismic signal processing method based on multi-feature fusion |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115220094B (en) | 2023-01-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhu et al. | Seismic signal augmentation to improve generalization of deep neural networks | |
| Saad et al. | Scalodeep: A highly generalized deep learning framework for real‐time earthquake detection | |
| Saad et al. | Earthquake detection and P-wave arrival time picking using capsule neural network | |
| Tang et al. | Identification of microseismic events in rock engineering by a convolutional neural network combined with an attention mechanism | |
| US11835674B1 (en) | Methods of analyzing cement integrity in annuli of a multiple-cased well using machine learning | |
| KR102094320B1 (en) | Method for improving image using reinforcement learning | |
| CN112799128B (en) | Method for seismic signal detection and seismic phase extraction | |
| CN111562612B (en) | Deep learning microseismic event identification method and system based on attention mechanism | |
| CN111985538A (en) | Small sample picture classification model and method based on semantic auxiliary attention mechanism | |
| Ku et al. | Attention-based convolutional neural network for earthquake event classification | |
| Rasmussen et al. | Automatic detection and classification of baleen whale social calls using convolutional neural networks | |
| CN115220094B (en) | Microseism event detection method based on BilSTM and attention mechanism | |
| Zheng et al. | Deep neural networks design and analysis for automatic phase pickers from three-component microseismic recordings | |
| CN116168295B (en) | Lithology remote sensing intelligent interpretation model establishment method and interpretation method | |
| Li et al. | Zoom out-and-in network with recursive training for object proposal | |
| CN103279744A (en) | Multi-scale tri-mode texture feature-based method and system for detecting counterfeit fingerprints | |
| Singh et al. | Automated Deep Learning based Disease Prediction Using Skin Health Records: Issues, Challenges and Future Directions | |
| CN115641473A (en) | Remote sensing image classification method based on CNN-self-attention mechanism hybrid architecture | |
| Duan et al. | Multi-scale convolutional neural network for SAR image semantic segmentation | |
| Wang et al. | Hyperspectral image classification based on non-local neural networks | |
| Bachay et al. | Hybrid Deep Learning Model Based on Autoencoder and CNN for Palmprint Authentication. | |
| Zhang et al. | Integrated processing method for microseismic signal based on deep neural network | |
| CN116299684B (en) | Novel microseismic classification method based on bimodal neurons in artificial neural network | |
| CN103646243A (en) | Digital-vein feature extraction method based on nonnegative-matrix factorization | |
| CN116068651A (en) | Seismic data classification method based on time attention mechanism |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |