Deep learning-based seismic geophysical monitoring image recognition and data anomaly detection method

Abstract

The invention belongs to the technical field of seismic geophysical monitoring, and provides a seismic geophysical monitoring image recognition and data anomaly detection method based on deep learning, which comprises the following steps: S1, acquiring seismic geophysical monitoring image data; s2, performing feature extraction on the image data by using a deep learning model; s3, identifying the image data according to the extracted features so as to detect data abnormity or other interference phenomena; s4, performing data analysis on an identification result; according to the method, the deep learning model is introduced, so that automatic feature extraction and efficient recognition of seismic geophysical monitoring image data are realized, and the accuracy and efficiency of image recognition are greatly improved; meanwhile, by using a preset anomaly detection algorithm, the method can accurately detect an abnormal value in the data and output an abnormal alarm in time, thereby providing powerful support for earthquake forecasting work and enhancing reliability and practicability of earthquake monitoring data.

Description

Technical Field

[0001] The present invention belongs to the technical field of earthquake geophysical monitoring, and specifically to an earthquake geophysical monitoring image recognition and data anomaly detection method based on deep learning. Background Art

[0002] In the field of earthquake geophysical monitoring, traditional image recognition and data anomaly detection methods have many shortcomings. Traditional image recognition methods often rely on manual feature extraction, which is not only inefficient, but also difficult to accurately capture the complex features in earthquake signals; at the same time, traditional data anomaly detection methods are usually based on statistical principles and manual recognition methods. For large-scale, high-dimensional earthquake geophysical monitoring data, the detection effect is not ideal, and it is easy to produce misjudgments and missed judgments; As a result, traditional image recognition and data anomaly detection methods can no longer meet the current needs of earthquake geophysical monitoring, so a new, efficient and accurate method is needed to solve these problems.

[0003] To this end, those skilled in the art have proposed a deep learning-based earthquake geophysical monitoring image recognition and data anomaly detection method to solve the problems raised by the background technology. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides an earthquake geophysical monitoring image recognition and data anomaly detection method based on deep learning to solve the many deficiencies in the image recognition and data anomaly detection methods in the prior art. Traditional image recognition methods often rely on manual feature extraction, which is not only inefficient, but also difficult to accurately capture the complex features in geophysical signals. At the same time, traditional data anomaly detection methods are usually based on statistical principles. For large-scale, high-dimensional earthquake geophysical monitoring data, their detection effect is not ideal and is prone to problems such as false positives and false negatives.

[0005] The deep learning-based earthquake geophysical monitoring image recognition and data anomaly detection method includes:

[0006] S1. Acquire earthquake geophysical monitoring image data;

[0007] S2. Extracting features from the image data using a deep learning model;

[0008] S3, identifying the image data based on the extracted features to detect anomalies or other interference phenomena;

[0009] S4, performing data analysis on the recognition results, and detecting outliers in the data using a preset anomaly detection algorithm, wherein the anomaly detection algorithm is based on the deviation of the output of the deep learning model from the normal data pattern;

[0010] S5. When an abnormal value is detected, an abnormal alarm is output.

[0011] Preferably, in step S1, a Kalman filter algorithm is used to remove noise from the acquired data to improve the accuracy and reliability of the data.

[0012] Preferably, in step S2, the deep learning model includes a convolutional neural network layer for automatically learning and extracting features of seismic signals;

[0013] The deep learning model also includes a recurrent neural network layer for processing time-series seismic geophysical monitoring data to improve the accuracy and timeliness of anomaly detection.

[0014] Preferably, a loss function of a deep learning model is introduced to achieve intelligent identification of structural damage.

[0015] Preferably, in step S3, the image data is identified based on the extracted features using a modal analysis algorithm to calculate the modal parameters of the structure (including modal frequency, modal damping ratio, modal vibration shape, etc.), understand the changing characteristics of the monitoring data, and provide reliable data for earthquake monitoring.

[0016] Preferably, the anomaly detection algorithm adopts an autoencoder structure and determines data anomalies by calculating reconstruction errors.

[0017] Preferably, the method further includes training the deep learning model, wherein the training data includes normal and abnormal image data with annotated abnormalities or other interference phenomena.

[0018] Preferably, the deep learning-based earthquake geophysical monitoring image recognition and data anomaly detection device uses the deep learning-based earthquake geophysical monitoring image recognition and data anomaly detection method, including:

[0019] The data acquisition module is used to acquire earthquake geophysical monitoring image data and use the Kalman filter algorithm to remove noise from the acquired data to improve the accuracy and reliability of the data;

[0020] A deep learning model module, comprising a convolutional neural network layer and a recurrent neural network layer, wherein the convolutional neural network layer is used to automatically learn and extract features of seismic signals, and the recurrent neural network layer is used to process time-series seismic geophysical monitoring data to improve the accuracy and timeliness of anomaly detection; the deep learning model module is also used to extract features from the image data;

[0021] The image recognition module is used to identify image data based on the features extracted by the deep learning model module to detect anomalies or other interference phenomena, and use the modal analysis algorithm to calculate the modal parameters of the structure to understand the dynamic characteristics of the monitoring data and provide reliable data for earthquake monitoring;

[0022] A data analysis module is used to perform data analysis on the output results of the image recognition module and preset an anomaly detection algorithm. The anomaly detection algorithm adopts an autoencoder structure and determines whether the data deviates from the normal data pattern by calculating the reconstruction error;

[0023] The abnormal alarm module is used to output an abnormal alarm when the data analysis module detects an abnormal value in the data;

[0024] A training module is used to train the deep learning model, and the training data includes normal and abnormal image data with annotated abnormalities or other phenomena.

[0025] A processor is configured to execute the above-mentioned deep learning-based earthquake geophysical monitoring image recognition and data anomaly detection method.

[0026] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above-mentioned deep learning-based earthquake geophysical monitoring image recognition and data anomaly detection method.

[0027] Compared with the prior art, the present invention has the following beneficial effects:

[0028] 1. The present invention realizes automatic feature extraction and recognition of earthquake geophysical monitoring image data by introducing a deep learning model, greatly improving the accuracy and efficiency of image recognition; the deep learning model can automatically learn the complex features in seismic signals, avoiding the tediousness and subjectivity of manual feature extraction.

[0029] 2. The anomaly detection algorithm proposed in the present invention is based on the deviation between the output of the deep learning model and the normal data pattern. It can accurately detect anomalies in the data and reduce the probability of false positives and negative negatives. At the same time, the algorithm also adopts an autoencoder structure to judge data anomalies by calculating the reconstruction error, further improving the accuracy of anomaly detection.

[0030] 3. The method of the present invention is not only suitable for processing static image data, but also can process time-series earthquake geophysical monitoring data, thereby improving the timeliness and accuracy of anomaly detection. By introducing a recurrent neural network layer, the method of the present invention can capture the dynamic features in time-series data, thus providing strong support for earthquake early warning and seismic resistance performance evaluation.

[0031] 4. The present invention also proposes a method for training a deep learning model, which is trained by annotating normal and abnormal image data with abnormalities or other interference phenomena, so that the deep learning model can better adapt to actual application scenarios, thereby improving the practicality and reliability of the method. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is a flow chart of the method for earthquake geophysical monitoring image recognition and data anomaly detection based on deep learning of the present invention;

[0033] Figure 2 This is a framework diagram of the deep learning-based earthquake geophysical monitoring image recognition and data anomaly detection device of the present invention. DETAILED DESCRIPTION

[0034] The following embodiments of the present invention are described in further detail in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0035] Embodiment: The present invention provides a method for earthquake geophysical monitoring image recognition and data anomaly detection based on deep learning, such as Figure 1 As shown, including:

[0036] S1. Acquire earthquake geophysical monitoring image data;

[0037] S2. Extracting features from the image data using a deep learning model;

[0038] S3, identifying the image data based on the extracted features to detect anomalies or other interference phenomena;

[0039] S4, performing data analysis on the recognition results, and detecting outliers in the data using a preset anomaly detection algorithm, wherein the anomaly detection algorithm is based on the deviation of the output of the deep learning model from the normal data pattern;

[0040] S5. When an abnormal value is detected, an abnormal alarm is output.

[0041] As can be seen from the above, by introducing the deep learning model, this method realizes the automatic feature extraction and efficient recognition of earthquake geophysical monitoring image data, greatly improving the accuracy and efficiency of image recognition. At the same time, using the preset anomaly detection algorithm, this method can accurately detect abnormal values ​​in the data and output abnormal alarms in time, providing strong support for earthquake prediction and enhancing the reliability and practicality of earthquake monitoring.

[0042] Furthermore, in step S1, a Kalman filter algorithm is used to remove noise from the acquired data to improve the accuracy and reliability of the data. The formula of the Kalman filter algorithm includes:

[0043]

[0044] in, is the state estimate, A and B are the system matrices, z k is the observation value, H is the observation matrix, K kis the Kalman gain.

[0045] From the above, we can see that by using the Kalman filter algorithm, we can effectively filter out the noise interference in the data, thereby ensuring the data quality of the subsequent deep learning model processing. This data preprocessing method not only optimizes the input of the deep learning model, but also further improves the performance of the entire earthquake geophysical monitoring image recognition and data anomaly detection method, laying a solid foundation for the subsequent abnormal or other interference phenomenon alarm output.

[0046] Furthermore, in step S2, the deep learning model includes a convolutional neural network layer for automatically learning and extracting features of seismic signals; the algorithm formula of the convolutional neural network layer includes:

[0047] y(i,j)=(K*I)(i,j)=∑ m ∑ n I(i+m,j+n)K(m,n);

[0048] Among them, I is the input image and K is the convolution kernel;

[0049] The deep learning model also includes a recurrent neural network layer for processing time-series seismic geophysical monitoring data to improve the accuracy and timeliness of anomaly detection. The formula of the recurrent neural network layer includes:

[0050] h t =σ(W hh h t-1 +W xh x t +b h );

[0051] Among them, h t is the hidden state, x t is the input.

[0052] As can be seen from the above, through the deep learning model including convolutional neural network layer and recurrent neural network layer, this method can automatically learn and extract complex features in seismic signals and effectively process time-series seismic geophysical monitoring data. The convolutional neural network layer uses a specific convolution kernel algorithm to deeply explore the key information in the input image, while the recurrent neural network layer captures the dynamic features in the time series data through its unique algorithm structure. The application of this deep learning model not only improves the accuracy and timeliness of anomaly detection, but also provides more accurate and comprehensive data support for earthquake early warning and seismic performance evaluation, further enhancing the practicality and reliability of earthquake monitoring.

[0053] Furthermore, the loss function of the deep learning model is introduced to realize intelligent identification of structural damage. The loss function algorithm formula includes:

[0054]

[0055] Among them, L is the loss function, N is the number of samples, and y i is the true label, is the predicted label.

[0056] From the above, we can see that in the earthquake geophysical monitoring image recognition and data anomaly detection method, the loss function of the deep learning model is introduced to realize the intelligent recognition of structural damage. Through the loss function algorithm, this method can quantify the difference between the real label and the predicted label, so as to accurately evaluate the prediction performance of the deep learning model. This mechanism not only helps to optimize the training process of the deep learning model and improve the model's recognition accuracy of seismic signals, but more importantly, it provides a powerful mathematical tool for realizing the intelligent recognition of data interference and anomalies. With the guidance of the loss function, this method can more accurately identify data anomalies and data changes caused by various interferences, providing a more reliable and scientific basis for earthquake monitoring.

[0057] Further, in step S3, the image data is identified according to the extracted features using a modal analysis algorithm to calculate the modal parameters of the structure (including modal frequency, modal damping ratio, modal vibration shape, etc.), understand the dynamic characteristics of the structure, and provide a basis for seismic performance evaluation. The formula of the modal analysis algorithm includes:

[0058] ([K]-w i 2 [M]){φ i} = 0;

[0059] Where [K] is the stiffness matrix, [M] is the mass matrix, and w i is the ith modal frequency, {φ i} is the i-th order mode shape.

[0060] From the above, we can see that by calculating the modal parameters of the structure, including modal frequency, modal damping ratio and modal vibration shape, this method can deeply understand the dynamic characteristics of the structure and provide a scientific basis for the evaluation of seismic performance. The use of modal analysis algorithms not only improves the accuracy of identifying anomalies or other interference phenomena, but also enables researchers to more accurately evaluate the response and performance of geophysical data under various interferences, thereby providing more powerful technical support for earthquake prediction and disaster reduction. The implementation of this step further enhances the practicality and scientificity of earthquake geophysical monitoring image recognition and data anomaly detection methods.

[0061] Furthermore, the anomaly detection algorithm adopts an autoencoder structure and determines data anomalies by calculating reconstruction errors.

[0062] From the above, it can be seen that in the earthquake geophysical monitoring image recognition and data anomaly detection methods, the anomaly detection algorithm using the autoencoder structure has significant beneficial effects. By calculating the reconstruction error, the algorithm can accurately determine whether the data is abnormal and effectively distinguish the difference between normal data and abnormal data. This anomaly detection method not only improves the accuracy and reliability of data anomaly detection, but also provides more accurate data support for earthquake early warning and seismic performance evaluation. The use of the autoencoder structure enables the anomaly detection algorithm to better adapt to large-scale, high-dimensional earthquake geophysical monitoring data, further improving the practicality and reliability of earthquake monitoring data.

[0063] Furthermore, it also includes training the deep learning model, and the training data includes normal and abnormal image data with annotated abnormalities or other interference phenomena.

[0064] As can be seen from the above, through the systematic training process, the deep learning model can learn the key features in the earthquake geophysical monitoring images, thereby significantly improving its ability to identify different interferences and anomalies. This training method based on a large amount of labeled data not only enhances the generalization ability of the model, enabling it to more accurately identify various complex interferences or anomalies, but also improves the practicality and reliability of the entire method. The trained deep learning model can analyze earthquake geophysical monitoring data more accurately, providing strong technical support for earthquake prediction work.

[0065] Furthermore, the effects of the earthquake geophysical monitoring image recognition and data anomaly detection method based on deep learning in the embodiment and the traditional image recognition and data anomaly detection method (comparative example) are compared to obtain the following table:

[0066]

[0067] As can be seen from the above table, compared with traditional image recognition and data anomaly detection methods, the deep learning-based earthquake geophysical monitoring image recognition and data anomaly detection method in the embodiment shows significant advantages in feature extraction, anomaly detection accuracy, processing data type, timeliness, generalization ability, practicality and reliability. This method can process earthquake geophysical monitoring data more efficiently and provide more accurate and reliable technical support for earthquake prediction.

[0068] Deep learning-based earthquake geophysical monitoring image recognition and data anomaly detection device, such as Figure 2 As shown, the above-mentioned deep learning-based earthquake geophysical monitoring image recognition and data anomaly detection method includes:

[0069] The data acquisition module is used to acquire earthquake geophysical monitoring image data and use the Kalman filter algorithm to remove noise from the acquired data to improve the accuracy and reliability of the data;

[0070] A deep learning model module, comprising a convolutional neural network layer and a recurrent neural network layer, wherein the convolutional neural network layer is used to automatically learn and extract features of seismic signals, and the recurrent neural network layer is used to process time-series seismic geophysical monitoring data to improve the accuracy and timeliness of anomaly detection; the deep learning model module is also used to extract features from the image data;

[0071] The image recognition module is used to identify image data based on the features extracted by the deep learning model module to detect anomalies or other interference phenomena, and use the modal analysis algorithm to calculate the modal parameters of the structure to understand the dynamic characteristics of the structure and provide a basis for seismic performance evaluation;

[0072] A data analysis module is used to perform data analysis on the output results of the image recognition module and preset an anomaly detection algorithm. The anomaly detection algorithm adopts an autoencoder structure and determines whether the data deviates from the normal data pattern by calculating the reconstruction error;

[0073] The abnormal alarm module is used to output an abnormal alarm when the data analysis module detects an abnormal value in the data;

[0074] A training module is used to train the deep learning model, and the training data includes normal and abnormal image data with annotated abnormalities or other interference phenomena.

[0075] The embodiment of the present application provides an electronic device, which is applicable to the above-mentioned deep learning-based earthquake geophysical monitoring image recognition and data anomaly detection method, including:

[0076] Memory, used to protect computer programs and data;

[0077] Processor, used to run system programs.

[0078] An embodiment of the present application provides a computer storage medium, which is suitable for the above-mentioned deep learning-based earthquake geophysical monitoring image recognition and data anomaly detection method, and performs hierarchical confidentiality management on the above-mentioned system and data in accordance with confidentiality management requirements.

[0079] Those skilled in the art will appreciate that the embodiments of the present application may be provided as a system or a computer program product. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0080] The present application is described with reference to the flowcharts and / or block diagrams of the devices (systems) and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0081] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0082] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0083] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0084] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0085] Computer readable media include permanent and non-permanent, removable and non-removable media, and can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include transitory media such as modulated data signals and carrier waves.

[0086] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, commodity or device including the elements.

[0087] The embodiments of the present invention are provided for the purpose of illustration and description. Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations of the present invention. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. A method for earthquake geophysical monitoring image recognition and data anomaly detection based on deep learning, characterized in that: include: S1. Acquire earthquake geophysical monitoring image data; S2. Extracting features from the image data using a deep learning model; S3, identifying the image data based on the extracted features to detect anomalies or other interference phenomena; S4, performing data analysis on the recognition results, and detecting outliers in the data using a preset anomaly detection algorithm, wherein the anomaly detection algorithm is based on the deviation of the output of the deep learning model from the normal data pattern; S5. When an abnormal value is detected, an abnormal alarm is output.

2. The method for earthquake geophysical monitoring image recognition and data anomaly detection based on deep learning as claimed in claim 1, characterized in that: In step S1, a Kalman filter algorithm is used to remove noise from the acquired data.

3. The method for earthquake geophysical monitoring image recognition and data anomaly detection based on deep learning as claimed in claim 1, characterized in that: In step S2, the deep learning model includes a convolutional neural network layer for automatically learning and extracting features of seismic signals; The deep learning model also includes a recurrent neural network layer for processing time-series seismic geophysical monitoring data.

4. The method for earthquake geophysical monitoring image recognition and data anomaly detection based on deep learning as claimed in claim 3, characterized in that: The loss function of the deep learning model is introduced to realize intelligent identification of structural damage.

5. The method for earthquake geophysical monitoring image recognition and data anomaly detection based on deep learning as claimed in claim 1, characterized in that: In step S3, the image data is identified based on the extracted features using a modal analysis algorithm.

6. The method for earthquake geophysical monitoring image recognition and data anomaly detection based on deep learning as claimed in claim 1, characterized in that: The anomaly detection algorithm adopts an autoencoder structure and determines data anomalies by calculating reconstruction errors.

7. The method for earthquake geophysical monitoring image recognition and data anomaly detection based on deep learning as claimed in claim 1, characterized in that: It also includes training the deep learning model, where the training data includes normal and abnormal image data with annotated abnormalities or other interference phenomena.

8. A device for earthquake geophysical monitoring image recognition and data anomaly detection based on deep learning, characterized in that: The method for earthquake geophysical monitoring image recognition and data anomaly detection based on deep learning according to any one of claims 1 to 7 comprises: The data acquisition module is used to acquire earthquake geophysical monitoring image data and use the Kalman filter algorithm to remove noise from the acquired data to improve the accuracy and reliability of the data; A deep learning model module, comprising a convolutional neural network layer and a recurrent neural network layer, wherein the convolutional neural network layer is used to automatically learn and extract features of seismic signals, and the recurrent neural network layer is used to process time-series seismic geophysical monitoring data to improve the accuracy and timeliness of anomaly detection; the deep learning model module is also used to extract features from the image data; The image recognition module is used to identify image data based on the features extracted by the deep learning model module to detect anomalies or other interference phenomena, and use the modal analysis algorithm to calculate the modal parameters of the structure to understand the changing characteristics of the monitoring data and provide reliable data for earthquake monitoring; A data analysis module is used to perform data analysis on the output results of the image recognition module and preset an anomaly detection algorithm. The anomaly detection algorithm adopts an autoencoder structure and determines whether the data deviates from the normal data pattern by calculating the reconstruction error; The abnormal alarm module is used to output an abnormal alarm when the data analysis module detects an abnormal value in the data; A training module is used to train the deep learning model, and the training data includes normal and abnormal image data with annotated abnormalities or other interference phenomena.

9. A processor, characterized in that: The method is configured to execute the deep learning-based earthquake geophysical monitoring image recognition and data anomaly detection method according to any one of claims 1 to 7.

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 earthquake geophysical monitoring image recognition and data anomaly detection based on deep learning described in any one of claims 1 to 7 is implemented.

Images