Mine micro-seismic event classification method and device, electronic equipment and storage medium

Abstract

The invention relates to a mine micro-seismic event classification method and device, electronic equipment and a storage medium, and belongs to the technical field of micro-seismic event monitoring, and the mine micro-seismic event classification comprises the steps: based on mine micro-seismic event real-time monitoring data, determining a first image and a second image, the first image being used for representing the relationship between signal amplitude and time, and the second image being used for representing the relationship between signal amplitude and time; the second image is used for representing the relationship between the signal frequency and time; and inputting the first image and the second image into a mine micro-seismic event classification model, and determining a classification result of the real-time monitoring data of the mine micro-seismic event based on the output of the mine micro-seismic event classification model. According to the method, while the classification accuracy of the mine micro-seismic events is ensured, the classification complexity of the mine micro-seismic events is reduced, and the reliability of the classification result of the mine micro-seismic events is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of microseismic event monitoring, and in particular to a method, device, electronic equipment and storage medium for classifying microseismic events in mines. Background Art

[0002] Microseismic monitoring plays a vital role in mine safety operations and disaster prevention. Mine microseisms are vibrations caused by the cracking or slipping of rock or ore body structures under stress. If these microseismic events are not handled in a timely manner, they may cause serious safety accidents such as rock bursts and collapses. Therefore, accurate and rapid identification and classification of microseismic signals are crucial to preventing these disasters. Traditional microseismic monitoring methods mainly rely on time domain and frequency domain analysis of seismic waves. These methods require professional geologists to manually interpret the characteristics of seismic waveforms for classification. This process is not only time-consuming, but also easily affected by personal experience and subjective judgment. It is difficult to adapt to the rapidly changing mining environment and the need to process large amounts of data.

[0003] Microseismic signals in mines are often accompanied by a large amount of background noise and interference from non-seismic signals, such as mechanical operations and electromagnetic interference. These interference signals are sometimes difficult to distinguish from microseismic signals in terms of waveform, which brings challenges to signal extraction and analysis. With the popularization of monitoring equipment and the increase in monitoring points, the amount of data generated is growing exponentially, and traditional manual or semi-automatic analysis methods can no longer meet the needs of real-time monitoring and rapid response. There are various types of microseismic events, ranging from tiny rock fractures to rock bursts that may cause large-scale damage, and different types of events require different countermeasures. Traditional methods are often limited in classification accuracy and reliability, especially when dealing with signals in complex or unknown environments.

[0004] Therefore, how to ensure the accuracy and reliability of the classification of microseismic events in mines has become a technical problem that needs to be solved urgently. Summary of the invention

[0005] In view of this, it is necessary to provide a method, device, electronic device and storage medium for classifying microseismic events in mines, so as to solve the problem that the current method for classifying microseismic events in mines is not accurate and reliable enough.

[0006] In order to solve the above problems, the present invention provides a method for classifying microseismic events in mines, comprising: Based on the real-time monitoring data of microseismic events in the mine, a first image and a second image are determined, wherein the first image is used to represent the relationship between signal amplitude and time, and the second image is used to represent the relationship between signal frequency and time; Inputting the first image and the second image into a mine microseismic event classification model, and determining a classification result of the mine microseismic event real-time monitoring data based on an output of the mine microseismic event classification model; Among them, the mine microseismic event classification model includes a residual network, a batch normalization network and a gating network set before the pooling layer, and an adaptive maximum pooling network set before the fully connected layer.

[0007] In a possible implementation, determining the first image and the second image based on the real-time monitoring data of microseismic events in a mine includes: Determining the first image based on the sampling frequency of the real-time monitoring data of the mine microseismic events; Based on the time information and amplitude information in the real-time monitoring data of microseismic events in the mine, a time-frequency analysis is performed on the real-time monitoring data of microseismic events in the mine to determine the second image.

[0008] In a possible implementation, determining the first image based on the sampling frequency of the real-time monitoring data of the mine microseismic events includes: Determine multiple sampling time points based on the sampling frequency and sampling duration of the real-time monitoring data of the mine microseismic events; The first image is determined based on the multiple sampling time points and the signal amplitudes at the multiple sampling time points.

[0009] In a possible implementation, the performing time-frequency analysis on the real-time monitoring data of microseismic events in the mine based on the time information and amplitude information in the real-time monitoring data of microseismic events in the mine to determine the second image includes: Based on the time information and amplitude information in the real-time monitoring data of microseismic events in the mine, S transformation is performed on the real-time monitoring data of microseismic events in the mine to determine the second image.

[0010] In a possible implementation, the mine microseismic event classification model includes a convolution layer, a feature adjustment layer, a pooling layer, an adaptive maximum pooling layer and a fully connected layer.

[0011] In a possible implementation, the feature adjustment layer includes a residual network, a batch normalization network and a gating network, and the adaptive maximum pooling layer includes an adaptive maximum pooling network.

[0012] In a possible implementation, the mine microseismic event classification model is trained based on historical monitoring data of mine microseismic events carrying real labels.

[0013] The present invention also provides a mine microseismic event classification device, comprising: A first determination module is used to determine a first image and a second image based on real-time monitoring data of microseismic events in a mine, wherein the first image is used to represent the relationship between signal amplitude and time, and the second image is used to represent the relationship between signal frequency and time; A second determination module is used to input the first image and the second image into a mine microseismic event classification model, and determine a classification result of the mine microseismic event real-time monitoring data based on an output of the mine microseismic event classification model; Among them, the mine microseismic event classification model includes a residual network, a batch normalization network and a gating network set before the pooling layer, and an adaptive maximum pooling network set before the fully connected layer.

[0014] The present invention also provides an electronic device, including a memory and a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the method for classifying microseismic events in mines as described above is implemented.

[0015] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the method for classifying microseismic events in a mine as described above is implemented.

[0016] The beneficial effects of the present invention are as follows: the mine microseismic event classification method, device, electronic device and storage medium provided by the present invention first convert the real-time monitoring data of mine microseismic events into image data, and obtain the relationship between the amplitude and frequency of the signal corresponding to the real-time monitoring data of mine microseismic events and time, which is more conducive to subsequent feature extraction, reduces the complexity of the data processing process, and thus improves the reliability of the classification results. Then, the mine microseismic event classification model is used to convert the real-time monitoring data of mine microseismic events into images for feature extraction, and classifies the microseismic events according to the feature information corresponding to different time nodes, thereby improving the accuracy of the microseismic event classification. While ensuring the accuracy of the mine microseismic event classification, the present invention reduces the complexity of the mine microseismic event classification and improves the reliability of the mine microseismic event classification results. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A schematic diagram of a flow chart of an embodiment of a method for classifying microseismic events in a mine provided by the present invention; Figure 2 A schematic diagram of a flow chart of an embodiment of a mine microseismic event classification process provided by the present invention; Figure 3 A schematic diagram of the structure of an embodiment of a feature adjustment layer provided by the present invention; Figure 4 A schematic diagram of the structure of an embodiment of a gated regulated convolutional neural network provided by the present invention; Figure 5 A schematic structural diagram of an embodiment of a mine microseismic event classification device provided by the present invention; Figure 6 A schematic structural diagram of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The preferred embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0019] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0020] In the description of the present invention, reference to an "embodiment" means that a particular feature, structure, or characteristic described in conjunction with the embodiment may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive of other embodiments. It is explicitly and implicitly understood by those skilled in the art that the described embodiments may be combined with other embodiments.

[0021] Microseismic monitoring plays a vital role in mine safety operations and disaster prevention. Mine microseisms are vibrations caused by the cracking or slipping of rock or ore body structures under stress. If these microseismic events are not handled in a timely manner, they may cause serious safety accidents such as rock bursts and collapses. Therefore, accurate and rapid identification and classification of microseismic signals are crucial to preventing these disasters. Traditional microseismic monitoring methods mainly rely on time domain and frequency domain analysis of seismic waves. These methods require professional geologists to manually interpret the characteristics of seismic waveforms for classification. This process is not only time-consuming, but also easily affected by personal experience and subjective judgment. It is difficult to adapt to the rapidly changing mining environment and the need to process large amounts of data.

[0022] Microseismic signals in mines are often accompanied by a large amount of background noise and interference from non-seismic signals, such as mechanical operations and electromagnetic interference. These interference signals are sometimes difficult to distinguish from microseismic signals in terms of waveform, which brings challenges to signal extraction and analysis. With the popularization of monitoring equipment and the increase in monitoring points, the amount of data generated is growing exponentially, and traditional manual or semi-automatic analysis methods can no longer meet the needs of real-time monitoring and rapid response. There are various types of microseismic events, ranging from tiny rock fractures to rock bursts that may cause large-scale damage, and different types of events require different countermeasures. Traditional methods are often limited in classification accuracy and reliability, especially when dealing with signals in complex or unknown environments.

[0023] In order to solve the problem in the prior art that the recognition accuracy of microseismic events and their mixed signals is limited, which affects the related disaster warning and prevention work, the present invention provides a method for classifying mine microseismic events.

[0024] The specific embodiments are described in detail below: A specific embodiment of the present invention discloses a method for classifying microseismic events in a mine, combining Figure 1 Come and see, Figure 1 The flowchart of an embodiment of a method for classifying microseismic events in a mine provided by the present invention includes step S101 and step S102, wherein: In step S101, based on the real-time monitoring data of microseismic events in a mine, a first image and a second image are determined, wherein the first image is used to represent the relationship between signal amplitude and time, and the second image is used to represent the relationship between signal frequency and time; In step S102, the first image and the second image are input into a mine microseismic event classification model, and a classification result of the mine microseismic event real-time monitoring data is determined based on the output of the mine microseismic event classification model; Among them, the mine microseismic event classification model includes a residual network, a batch normalization network and a gating network set before the pooling layer, and an adaptive maximum pooling network set before the fully connected layer.

[0025] During implementation, first, real-time monitoring data of microseismic events in a mine over a period of time can be obtained, and then the first image and the second image can be determined based on these real-time monitoring data. The first image can be used to represent the relationship between signal amplitude and time, and the second image can be used to represent the relationship between signal frequency and time, that is, through the first image and the second image, the relationship between the amplitude and frequency of the signal corresponding to the real-time monitoring data of microseismic events in the mine and time can be determined, thereby providing a data basis for subsequent feature extraction and microseismic event classification.

[0026] After determining the first image and the second image, the first image and the second image can be input into a mine microseismic event classification model. The mine microseismic event classification model can extract features and classify the first image and the second image. According to the output of the mine microseismic event classification model, the classification result of the real-time monitoring data of mine microseismic events can be determined.

[0027] The mine microseismic event classification model may include a residual network, a batch normalization network and a gating network set before the pooling layer, and an adaptive maximum pooling network set before the fully connected layer. It can extract and identify feature information from inputs of different sizes. Since the feature information corresponding to different microseismic events is also different, the microseismic events can be classified according to the feature information corresponding to different time nodes.

[0028] The mine microseismic event classification method provided by the present invention can be applied to mine microseismic event classification scenarios, and can also be applied to microseismic event classification scenarios in other geological exploration activities, and the present invention does not make specific limitations on this.

[0029] Compared with the prior art, the mine microseismic event classification method provided in this embodiment first converts the real-time monitoring data of mine microseismic events into image data, and obtains the relationship between the amplitude and frequency of the signal corresponding to the real-time monitoring data of mine microseismic events and time, which is more conducive to subsequent feature extraction, reduces the complexity of the data processing process, and thus improves the reliability of the classification results. Then, the mine microseismic event classification model is used to convert the real-time monitoring data of mine microseismic events to obtain an image for feature extraction, and classifies the microseismic events according to the feature information corresponding to different time nodes, thereby improving the accuracy of the microseismic event classification. While ensuring the accuracy of the mine microseismic event classification, the present invention reduces the complexity of the mine microseismic event classification and improves the reliability of the mine microseismic event classification results.

[0030] Exemplarily, the determining of the first image and the second image based on the real-time monitoring data of microseismic events in a mine includes: Determining the first image based on the sampling frequency of the real-time monitoring data of the mine microseismic events; Based on the time information and amplitude information in the real-time monitoring data of microseismic events in the mine, a time-frequency analysis is performed on the real-time monitoring data of microseismic events in the mine to determine the second image.

[0031] Specifically, when determining the first image and the second image based on the real-time monitoring data of microseismic events in a mine, the first image can be determined based on the sampling frequency of the real-time monitoring data of microseismic events in a mine, and based on the time information and amplitude information in the real-time monitoring data of microseismic events in a mine, a time-frequency analysis (such as waveform transformation and other means) is performed on the real-time monitoring data of microseismic events in a mine to determine the second image, thereby converting the traditional time-frequency analysis task into an image recognition task, thereby improving the reliability of the microseismic event classification results.

[0032] Exemplarily, the determining the first image based on the sampling frequency of the real-time monitoring data of the mine microseismic events includes: Determine multiple sampling time points based on the sampling frequency and sampling duration of the real-time monitoring data of the mine microseismic events; The first image is determined based on the multiple sampling time points and the signal amplitudes at the multiple sampling time points.

[0033] Specifically, when determining the first image according to the sampling frequency of the real-time monitoring data of microseismic events in a mine, multiple sampling time points may be determined according to the sampling frequency and sampling duration of the real-time monitoring data of microseismic events in a mine.

[0034] For example, the interval of each sampling time point can be determined according to the following formula;

[0035] in, represents the interval of each sampling time point, Indicates the sampling duration, Indicates the sampling frequency.

[0036] After the multiple sampling time points are determined, an amplitude-time image may be generated according to the signal amplitudes at the multiple sampling time points to obtain a first image.

[0037] Exemplarily, the performing time-frequency analysis on the real-time monitoring data of microseismic events in the mine based on the time information and amplitude information in the real-time monitoring data of microseismic events in the mine to determine the second image includes: Based on the time information and amplitude information in the real-time monitoring data of microseismic events in the mine, S transformation is performed on the real-time monitoring data of microseismic events in the mine to determine the second image.

[0038] Specifically, when performing time-frequency analysis on the real-time monitoring data of microseismic events in a mine according to the time information and amplitude information in the real-time monitoring data of microseismic events in a mine, the real-time monitoring data of microseismic events in a mine can be subjected to S transformation according to the time information and amplitude information in the real-time monitoring data of microseismic events in a mine, thereby obtaining a second image. The S transformation can locate the time and frequency content of the signal, and is suitable for analyzing non-stationary signals, and the signal corresponding to the microseismic event in a mine is a typical non-stationary signal.

[0039] For example, for discrete microseismic event signals , its S transform can be expressed as:

[0040] in, Indicates the time index and frequency The S-transform result at represents the signal amplitude, Indicates a time node. Indicates the number of samples.

[0041] Exemplarily, the mine microseismic event classification model includes a convolution layer, a feature adjustment layer, a pooling layer, an adaptive maximum pooling layer and a fully connected layer.

[0042] Specifically, the mine microseismic event classification model includes network structures such as convolutional layer, feature adjustment layer, pooling layer, adaptive maximum pooling layer and fully connected layer. The feature adjustment layer can deepen the network by increasing nonlinearity and controlling the information flow while keeping the size of the feature map unchanged. The adaptive maximum pooling layer can ensure that the size of the output feature vector of the microseismic event is consistent after multi-layer processing and ensure the significance of the feature, so that important feature information can be captured even in inputs of different sizes.

[0043] Exemplarily, the feature adjustment layer includes a residual network, a batch normalization network and a gating network, and the adaptive maximum pooling layer includes an adaptive maximum pooling network.

[0044] Specifically, the feature adjustment layer may include a residual network, a batch normalization network, and a gating network, and more efficient parallel computing capabilities are achieved by adopting convolution and gating mechanisms. The adaptive maximum pooling layer includes an adaptive maximum pooling network, which can dynamically adjust the size of the pooling window, allowing the network to receive input signals of any length and output a fixed-length feature vector.

[0045] Exemplarily, the mine microseismic event classification model is trained based on historical monitoring data of mine microseismic events carrying real labels.

[0046] Specifically, the mine microseismic event classification model can be trained based on the historical monitoring data of mine microseismic events carrying real labels. During training, the historical monitoring data of mine microseismic events can be converted into image data first, and then the image data can be input into the mine microseismic event classification model. The model parameters are adjusted according to the difference between the model output and the real label until the model classification effect meets the requirements.

[0047] The following is a specific application scenario to better illustrate the technical solution of the present invention: Combination Figure 2 Come and see, Figure 2 This is a flow chart of an embodiment of a mine microseismic event classification process provided by the present invention. The process mainly includes the following steps: 1. Image generation: Generate image datasets based on the collected data.

[0048] A waveform diagram is generated according to the time-amplitude information in the original data, and an S-transform is performed according to the time-amplitude information in the original data to generate an S-transform spectrum diagram.

[0049] The microseismic event waveform diagram is composed of the microseismic event time parameter and the corresponding amplitude parameter, and is generally drawn as a line graph in actual image generation.

[0050] For discrete microseismic event signals , according to its sampling frequency , the actual time corresponding to each sample It can be calculated by the following formula:

[0051] With time and signal The value of can generate the waveform diagram of microseismic events.

[0052] Stockwell transform, or S transform, is a time-frequency analysis technique that combines the advantages of short-time Fourier transform (STFT) and continuous wavelet transform (CWT). It is an extension of the idea of ​​continuous wavelet transform and uses a movable and retractable Gaussian window function. The special feature of the S transform is that it has good frequency-related resolution and is directly related to the Fourier transform. In signal analysis, a stationary signal is only an idealized mathematical model. Although the Fourier transform contains overall spectral information, for most actual signals, the spectral components of the time series are highly correlated with time, such as seismic wave signals. The S transform outputs a two-dimensional time-frequency representation, where the first dimension of the output represents time and the second dimension represents frequency. The S transform can locate the time and frequency content of the signal and is suitable for analyzing non-stationary signals. For discrete microseismic event signals , its S transform can be expressed as:

[0053] in, Is the time index and frequency The S-transform result at .

[0054] 2. Network composition: Combining the long short-term memory network (LSTM) and the residual network (ResNet) to construct a gated regulated convolutional neural network.

[0055] Combination Figure 3 Come and see, Figure 3 This is a schematic diagram of the structure of an embodiment of the feature adjustment layer provided by the present invention. Compared with the traditional LSTM, this network layer (hereinafter referred to as the GFM module) achieves more efficient parallel computing capabilities by adopting convolution and gating mechanisms. At the same time, by improving the combination of residual connections and output gates, it not only solves the information flow and gradient problems in the deep network, but also increases the dynamic adjustment of feature importance, thereby maintaining effective information flow while enhancing the expressiveness of features.

[0056] The forward propagation calculation process of the GFM module can be expressed as:

[0057] Among them, Conv1 represents the first convolution operation and BN1 represents the first batch normalization operation.

[0058]

[0059] Similarly, Conv2 and BN2 represent the second convolution and normalization operations, respectively.

[0060]

[0061] It is the Sigmoid activation function, which is used to generate a gating signal between 0 and 1. , the final output is the gating signal and Element-wise product of :

[0062] The purpose of the entire GFM module is to deepen the network by adding nonlinearity and the ability to control information flow while keeping the feature map size unchanged.

[0063] Adaptive max pooling technology allows the network to receive input signals of any length and output feature vectors of fixed length by dynamically adjusting the size of the pooling window. Specifically, for any length Input signal , the adaptive max pooling layer converts it to a fixed length The output vector ,in It is determined by the downstream structure of the network and is a pre-set target length. This process can be expressed as:

[0064] in, Represents the output vector The elements, Indicates input signal In the By selecting the maximum value in each window, adaptive max pooling not only ensures that the size of the output feature vector of microseismic events is consistent after multi-layer processing, but also ensures the saliency of the features, and can capture the most important feature information even in inputs of different sizes.

[0065] 3. Model training and capability evaluation: Use image datasets to train gated convolutional neural network models and evaluate model classification performance.

[0066] The model is trained according to the image dataset to obtain a gated regulated convolutional neural network model.

[0067] According to the number and frequency of sampling points, microseismic events are divided into different intervals. The gated convolutional neural network model is called to evaluate the model classification performance from three dimensions: F1 score, accuracy and recall rate.

[0068] The gated convolutional neural network architecture adopts a multi-level, staged feature extraction and processing method, combining the GFM module and convolution layer, and realizes accurate and automatic classification of microseismic events through appropriate pooling layers and fully connected layers.

[0069] In the gated convolutional neural network, complex microseismic event signal data is mapped to a high-dimensional feature space through convolution and instance normalization, ensuring that subtle differences in microseismic event signals can be effectively captured, and deep feature extraction and adjustment are performed based on the characteristics of microseismic event signals, thereby better capturing information that is important for classification decisions. Figure 4 Come and see, Figure 4 The schematic diagram of the structure of an embodiment of the gated regulated convolutional neural network provided by the present invention. The specific structural parameters of the gated regulated convolutional neural network are shown in Table 1.

[0070] Table 1: Gated convolutional neural network model structure parameters

[0071] The results of each evaluation index of the model are shown in Table 2.

[0072] Table 2: Evaluation indicators of gated convolutional neural network model

[0073] According to the model test results, the performance of the gated regulated convolutional neural network model on each signal category was comprehensively evaluated and analyzed from three dimensions: accuracy, recall rate and F1 score.

[0074] First, in terms of rock fracture signal recognition, the model has an accuracy of 0.9311, a recall of 0.9600, and an F1 score of 0.9453. The F1 score is at a high level, indicating that the gated convolutional neural network model has achieved a good balance in the task of rock fracture signal recognition, and can maintain high accuracy and coverage at the same time.

[0075] For the recognition of blasting signals, the accuracy of the gated convolutional neural network model was 91.43%, slightly lower than that of rock fracture signals, which means that the model mistakenly identified some non-blasting signals as blasting signals. The recall rate was 95.00%, indicating that the model was also quite accurate in identifying actual blasting signals, and not too many blasting signals were missed. The F1 score was 93.18%, showing that the model also performed well in blasting signal classification, but there is still room for improvement, especially in reducing false alarms.

[0076] The gated convolutional neural network has an accuracy rate of 99.60% for electromagnetic signal recognition, the highest among all categories, indicating that the model rarely misclassifies other types of signals as electromagnetic signals. The recall rate is 99.30%, and almost all electromagnetic signals are correctly detected, with a very low missed detection rate. The F1 score is 99.45%, showing that the model has excellent performance in classifying electromagnetic signals.

[0077] In addition, for the recognition of mechanical operation signals, the model's accuracy is 0.9860, but the recall rate is slightly lower, which is 0.9110, resulting in an F1 score of 0.9470, which is lower than other categories, meaning that a certain proportion of mechanical operation signals are misclassified as other types of signals.

[0078] Finally, in the task of human voice signal recognition, the gated regulated convolutional neural network model still achieved extremely outstanding results, with an accuracy of 0.9782, and performed well in all categories, indicating that the model did a good job in distinguishing human voice signals from other signals. The recall rate was 0.9870, which means that the model can accurately capture most human voice signals. The corresponding F1 score is as high as 0.9826, almost reaching the optimal level. These data fully demonstrate the model's excellent recognition ability for human voice signals.

[0079] In summary, the gated convolutional neural network model performs well in a variety of signal classification tasks. Although the recognition of mechanical operation signals is slightly lower than that of other categories, in general, the model has demonstrated a strong ability to classify and recognize five types of microseismic events.

[0080] In the same training environment, the present invention also keeps the number of main feature extraction modules and training parameters consistent with the gated convolutional neural network, selects CNN, ResNet and ConvLSTM networks for comparison, and evaluates the model classification performance from three dimensions: F1 score, accuracy and recall rate.

[0081] In contrast, the gated regulated convolutional neural network model performed best in the classification task of five types of signals. The average F1 score of the gated regulated convolutional neural network was 0.960, significantly higher than the CNN (0.867), ResNet (0.901) and ConvLSTM (0.923) models. The F1 scores of the gated regulated convolutional neural network in all categories were higher than those of the other three models, especially in the classification of mechanical operation signals, with an F1 value of 0.947, which greatly surpassed the CNN, ResNet and other models. The gated regulated convolutional neural network has an outstanding ability to recognize two relatively simple categories, electromagnetic signals and human voice signals, with F1 values ​​of 0.994 and 0.983 respectively, showing its strong feature extraction ability. Even in the difficult-to-identify rock fracture signals and blasting signals, the F1 scores of the gated regulated convolutional neural network reached 0.945 and 0.932 respectively, which are much higher than the other three comparison models.

[0082] The present invention can effectively transform traditional time-frequency analysis into an image recognition task, and utilizes modern deep learning technology to significantly improve the automatic recognition accuracy and processing efficiency of microseismic events, especially when processing large-scale and complex data. In addition, the present invention combines residual connections and gating mechanisms to complete a more sophisticated feature extraction process, and is conducive to the construction of a deep network, further improving the accuracy of identifying microseismic events in unknown or changing mining environments.

[0083] The embodiment of the present invention also provides a mine microseismic event classification device, combined with Figure 5 Come and see, Figure 5 The structure diagram of an embodiment of a mine microseismic event classification device provided by the present invention is as follows. The mine microseismic event classification device 500 includes: A first determination module 501 is used to determine a first image and a second image based on real-time monitoring data of microseismic events in a mine, wherein the first image is used to represent the relationship between signal amplitude and time, and the second image is used to represent the relationship between signal frequency and time; A second determination module 502 is used to input the first image and the second image into a mine microseismic event classification model, and determine a classification result of the mine microseismic event real-time monitoring data based on the output of the mine microseismic event classification model; Among them, the mine microseismic event classification model includes a residual network, a batch normalization network and a gating network set before the pooling layer, and an adaptive maximum pooling network set before the fully connected layer.

[0084] The specific implementation of each module of the mine microseismic event classification device can refer to the description of the above-mentioned mine microseismic event classification method, and has similar beneficial effects, which will not be repeated here.

[0085] It should be noted that the mine microseismic event classification device can be set in a mine operation monitoring center or in a detection workbench of a mine microseismic event monitoring personnel, and the present invention does not make specific limitations on this.

[0086] The embodiment of the present invention further provides an electronic device, Figure 6 Come and see, Figure 6 This is a structural diagram of an embodiment of an electronic device provided by the present invention. The electronic device 600 includes a processor 601, a memory 602, and a computer program stored in the memory 602 and executable on the processor 601. When the processor 601 executes the program, the method for classifying microseismic events in mines as described above is implemented.

[0087] As a preferred embodiment, the electronic device 600 further includes a display 603 for displaying the processor 601 executing the above-mentioned mine microseismic event classification method.

[0088] Exemplarily, the computer program may be divided into one or more modules / units, one or more modules / units are stored in the memory 602, and executed by the processor 601 to complete the present invention. One or more modules / units may be a series of computer program instruction segments that can complete specific functions, and the instruction segments are used to describe the execution process of the computer program in the electronic device 600. For example, the computer program may be divided into the first determination module 501 and the second determination module 502 in the above embodiment, and the specific functions of each module are as described above, and are not repeated here one by one.

[0089] The electronic device 600 may be a desktop computer, a notebook, a PDA or a smart phone with an adjustable camera module.

[0090] Among them, the processor 601 may be an integrated circuit chip with signal processing capabilities. The above-mentioned processor 601 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The disclosed methods, steps and logic block diagrams in the embodiments of the present invention can be implemented or executed. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0091] The memory 602 may be, but is not limited to, a random access memory (RAM), a read only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable read-only memory (EEPROM), etc. The memory 602 is used to store programs, and the processor 601 executes the program after receiving an execution instruction. The process definition method disclosed in any of the above embodiments of the present invention can be applied to the processor 601, or implemented by the processor 601.

[0092] The display 603 may be an LCD display screen or an LED display screen, for example, a display screen on a mobile phone.

[0093] Understandably, Figure 6 The structure shown is only a schematic diagram of a structure of the electronic device 600. The electronic device 600 may also include Figure 6 More or fewer components as shown. Figure 6 Each component shown in the figure can be implemented by hardware, software or a combination thereof.

[0094] The electronic device provided according to the above embodiment of the present invention can be implemented with reference to the specific description of the method for classifying microseismic events in mines as described above according to the present invention, and has similar beneficial effects as the method for classifying microseismic events in mines as described above, which will not be described in detail here.

[0095] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the method for classifying microseismic events in a mine as described above is implemented.

[0096] Generally speaking, the computer instructions for implementing the method of the present invention can be carried by any combination of one or more computer-readable storage media. Non-transitory computer-readable storage media can include any computer-readable media except for the signal itself that is temporarily propagating.

[0097] Computer-readable storage media may be, for example, but not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more conductors, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium may be any tangible medium that contains or stores a program that may be used by or in conjunction with an instruction execution system, apparatus, or device.

[0098] Computer program codes for performing the operations of the present invention may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages, in particular, Python suitable for neural network computing and platform frameworks based on TensorFlow, PyTorch, etc. may be used. The program code may be executed entirely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0099] Those skilled in the art will appreciate that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0100] The present invention discloses a method, device, electronic device and storage medium for classifying microseismic events in a mine. Firstly, real-time monitoring data of microseismic events in a mine is converted into image data to obtain the relationship between the amplitude and frequency of the signal corresponding to the real-time monitoring data of microseismic events in a mine and time, which is more conducive to subsequent feature extraction, reduces the complexity of the data processing process, and further improves the reliability of the classification result. Then, a classification model for microseismic events in a mine is used to convert the real-time monitoring data of microseismic events in a mine to obtain an image for feature extraction, and classifies microseismic events according to feature information corresponding to different time nodes, thereby improving the accuracy of microseismic event classification. While ensuring the accuracy of microseismic event classification in a mine, the present invention reduces the complexity of microseismic event classification in a mine and improves the reliability of the classification result of microseismic events in a mine.

[0101] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for classifying microseismic events in mines, characterized in that: include: Based on the real-time monitoring data of microseismic events in the mine, a first image and a second image are determined, wherein the first image is used to represent the relationship between signal amplitude and time, and the second image is used to represent the relationship between signal frequency and time; Inputting the first image and the second image into a mine microseismic event classification model, and determining a classification result of the mine microseismic event real-time monitoring data based on an output of the mine microseismic event classification model; Among them, the mine microseismic event classification model includes a residual network, a batch normalization network and a gating network set before the pooling layer, and an adaptive maximum pooling network set before the fully connected layer.

2. The method for classifying microseismic events in mines according to claim 1, characterized in that: The determining of the first image and the second image based on the real-time monitoring data of the mine microseismic events includes: Determining the first image based on the sampling frequency of the real-time monitoring data of the mine microseismic events; Based on the time information and amplitude information in the real-time monitoring data of microseismic events in the mine, a time-frequency analysis is performed on the real-time monitoring data of microseismic events in the mine to determine the second image.

3. The method for classifying microseismic events in mines according to claim 2, characterized in that: The determining of the first image based on the sampling frequency of the real-time monitoring data of the mine microseismic events includes: Determine multiple sampling time points based on the sampling frequency and sampling duration of the real-time monitoring data of the mine microseismic events; The first image is determined based on the multiple sampling time points and the signal amplitudes at the multiple sampling time points.

4. The method for classifying microseismic events in mines according to claim 2, characterized in that: The step of performing time-frequency analysis on the real-time monitoring data of microseismic events in the mine based on the time information and amplitude information in the real-time monitoring data of microseismic events in the mine to determine the second image includes: Based on the time information and amplitude information in the real-time monitoring data of microseismic events in the mine, S transformation is performed on the real-time monitoring data of microseismic events in the mine to determine the second image.

5. The method for classifying microseismic events in mines according to claim 1, characterized in that: The mine microseismic event classification model includes a convolution layer, a feature adjustment layer, a pooling layer, an adaptive maximum pooling layer and a fully connected layer.

6. The method for classifying microseismic events in mines according to claim 5, characterized in that: The feature adjustment layer includes a residual network, a batch normalization network and a gating network, and the adaptive maximum pooling layer includes an adaptive maximum pooling network.

7. The method for classifying microseismic events in a mine according to any one of claims 1 to 6, characterized in that: The mine microseismic event classification model is trained based on historical monitoring data of mine microseismic events carrying real labels.

8. A mine microseismic event classification device, characterized in that: include: A first determination module is used to determine a first image and a second image based on real-time monitoring data of microseismic events in a mine, wherein the first image is used to represent the relationship between signal amplitude and time, and the second image is used to represent the relationship between signal frequency and time; A second determination module is used to input the first image and the second image into a mine microseismic event classification model, and determine a classification result of the mine microseismic event real-time monitoring data based on an output of the mine microseismic event classification model; Among them, the mine microseismic event classification model includes a residual network, a batch normalization network and a gating network set before the pooling layer, and an adaptive maximum pooling network set before the fully connected layer.

9. An electronic device, characterized in that: The invention comprises a memory and a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the method for classifying microseismic events in a mine according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by a processor, the method for classifying microseismic events in a mine as claimed in any one of claims 1 to 7 is implemented.

Images