Seismic data intelligent quality control method and device based on machine learning

Abstract

The invention discloses a seismic data intelligent quality control method and device based on machine learning, and belongs to the technical field of seismic data processing. The method comprises the following steps: S1, calculating single-channel attributes with correlation and sensitivity according to input seismic data and a processing purpose, and organizing according to different data fields; s2, performing feature extraction on attributes of different data fields by using a deep learning network to generate a feature matrix; s3, classifying the feature matrix by using a clustering analysis algorithm to realize classification of seismic data; and S4, performing targeted data processing by using the classification result of the seismic data to realize quality control of the seismic data. By introducing a deep learning network and an unsupervised learning technology in machine learning, automatic intelligent quality control is realized, the seismic data processing efficiency is remarkably improved, and the error rate can be effectively reduced. The method is suitable for quality control in seismic data processing.

Description

Technical Field

[0001] The present invention belongs to the technical field of seismic data processing, and relates to a quality control method in the process of seismic data processing. Specifically, it is an intelligent quality control method and device for seismic data based on machine learning. Background Art

[0002] Seismic exploration is a key technology in modern geological science and resource exploration. By collecting and processing the propagation of seismic waves in different geological structures, the underground geological structure can be effectively interpreted. This technology plays a crucial role in the exploration of oil, natural gas, mineral resources, and groundwater. Moreover, seismic exploration also provides extremely important data support for assessing geological disasters such as earthquakes and landslides.

[0003] With the progress of technology and the increasing exploration requirements, the scale of seismic data acquisition has been continuously expanding, as evidenced by the popularization of the "two-wide and one-high" seismic acquisition method. This method has increased the trace density and the area of the work area, causing the seismic data volume of a single work area to rapidly grow from the TB level to the PB level. The sharp increase in data volume has posed unprecedented challenges to the computational efficiency of the seismic data processing flow, and the quality control (QC) link is particularly crucial. Good data quality control can not only ensure the accuracy of the data but also effectively improve the efficiency and refinement of subsequent data processing and interpretation.

[0004] Traditional seismic data quality control methods mainly rely on manual inspection. However, when faced with large-scale data at the PB level, this method is inadequate. Manual inspection is not only time-consuming but also easily affected by the skills and experience of data processing personnel, and the results often lack consistency. Although automated quality control technologies have made certain progress, such as rule-based anomaly detection, these methods usually cannot flexibly handle complex seismic data. This is particularly prominent in the process of transitioning from land exploration to more complex marine exploration. With the expansion of the exploration area and the increase in data complexity, the task of quality control has become increasingly complex and time-consuming.

[0005] In response to the challenges faced by existing quality control methods, existing research has shown that machine learning-based technologies can be used for automated data quality detection. However, most applications still rely on supervised learning methods, which require a large amount of labeled data and are not suitable for quickly analyzing large-scale seismic data. Summary of the Invention

[0006] The purpose of the present invention is to provide an intelligent quality control method and device for seismic data based on machine learning to solve the problem that there is currently no efficient quality control method applicable to large-scale seismic data.

[0007] To achieve the above purpose, the technical solution adopted by the present invention is: An intelligent quality control method for seismic data based on machine learning, comprising the following steps: S1. Calculation of seismic data attributes: Calculate single-trace attributes with relevance and sensitivity according to the input seismic data and processing purposes, and organize them according to different data domains; S2. Feature extraction of seismic data attributes: Use a deep learning network to extract features from the attributes in different data domains to generate a feature matrix; S3. Classification of seismic data: Use a clustering analysis algorithm to classify the feature matrix to achieve the classification of seismic data; S4. Quality control of seismic data: Apply the classification results of seismic data for targeted data processing to achieve the quality control of seismic data.

[0008] As a limitation, when there are multiple single-trace attributes with relevance and sensitivity calculated in step S1, the principal component analysis method is used to achieve data dimensionality reduction.

[0009] As another limitation, it also includes preprocessing the seismic data before step S1.

[0010] As a third limitation, the seismic data includes prestack data or poststack data.

[0011] As a fourth limitation, the attributes include statistical domain attributes, spectral domain attributes, and time domain attributes.

[0012] As a further limitation, the statistical domain attributes include maximum value, minimum value, mean value, median value, skewness, kurtosis, and time difference; the spectral domain attributes include Fourier transform coefficients, wavelet transform parameters, and spectral distance; the time domain attributes include autocorrelation function, differential mean value, and entropy.

[0013] As a fifth limitation, the data domain includes shot domain or geophone domain.

[0014] As a sixth limitation, the deep learning network includes a convolutional neural network (CNN).

[0015] As a seventh limitation, the clustering analysis algorithm includes the K-means algorithm.

[0016] The present invention also provides an intelligent quality control device for seismic data based on machine learning, comprising the following modules connected in sequence: Seismic data attribute calculation module: Used to calculate single-trace attributes with relevance and sensitivity according to the input seismic data and processing purposes, and organize them according to different data domains; Seismic data attribute feature extraction module: Used to extract features from the attributes in different data domains by using a deep learning network to generate a feature matrix; Seismic data classification module: Classify the feature matrix using a clustering analysis algorithm to achieve the classification of seismic data; Seismic data quality control module: Apply the classification results of seismic data for targeted data processing to achieve the quality control of seismic data.

[0017] The present invention also provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the intelligent quality control method for seismic data based on machine learning described above is implemented.

[0018] The present invention also provides a computer-readable storage medium storing a computer program for executing the intelligent quality control method for seismic data based on machine learning described above.

[0019] Due to the adoption of the above technical solutions, compared with the prior art, the technical progress achieved by the present invention is as follows: An intelligent quality control method and device for seismic data based on machine learning provided by the present invention can achieve automatic intelligent quality control in the process of seismic data processing, significantly improve the efficiency of seismic data processing, accurately identify abnormal nodes and other quality problems (such as noise, instrument response anomalies, and positioning errors) in the data, effectively reduce the error rate, and reduce misoperations on normal data by introducing deep learning networks and unsupervised learning techniques in machine learning. Moreover, it has good adaptability and scalability, can adapt to the more complex and larger-scale seismic data processing requirements in the future, thereby improving the quality and efficiency of the entire seismic exploration project. Description of the Drawings

[0020] Figure 1 It is a flowchart of the method in Embodiment 1; Figure 2 It is the display result of pre-stack seismic data in GeoEast software in Embodiment 1; Figure 3 It is a node classification result diagram in Embodiment 1, where Figure 3 (a) is the first type of node, Figure 3 (b) is the second type of node, Figure 3 (c) is the third type of node; Figure 4 It is a schematic structural diagram of the device in Embodiment 2.

[0021] 1. Seismic data attribute calculation module; 2. Seismic data attribute feature extraction module; 3. Seismic data classification module; 4. Seismic data quality control module. Detailed Embodiments

[0022] The present invention will be further described in detail below through specific embodiments. It should be understood that the described embodiments are only used to explain the present invention and do not limit the present invention.

[0023] Embodiment 1 This embodiment discloses an intelligent quality control method for seismic data based on machine learning. The flowchart of the method is as Figure 1 shown, and this method is applied to the quality control in the secondary positioning process of a certain ocean bottom node (OBN) seismic geophone point.

[0024] In offshore exploration operations, nodes are placed on the sea surface according to the designed coordinates using GPS and other positioning systems (primary positioning). Due to the influence of factors such as ocean currents, tides, ship speed, and sea waves, the actual sinking position of the nodes often deviates from the predetermined position. Even if the nodes are initially placed accurately, fishing boat activities and ocean climate changes during exploration may also cause the node positions to shift. This position shift not only affects the acquisition quality of seismic data but also increases the difficulty of subsequent seismic data processing.

[0025] By using the method provided in this embodiment, based on the pre-stack seismic data, different types of nodes can be accurately identified through node classification in the geophone point domain, and then targeted secondary positioning processing and parameter selection can be implemented, thereby improving the efficiency and accuracy of secondary positioning. The specific steps are as follows: S0. Seismic data preprocessing All kinds of seismic data collected (including pre-stack data or post-stack data, which is pre-stack data in this embodiment, specifically the pre-stack seismic data of OBN) are input into a specific seismic processing and interpretation software (GeoEast software is used in this embodiment), and the result is as Figure 2 shown.

[0026] During this process, according to the limitations of the equipment and environmental conditions used during data acquisition, the seismic data can be preprocessed (such as checking the acquisition system, removing damaged data channels, noise suppression, etc.). At the same time, it is also possible not to perform preprocessing and directly proceed to the next step. In this embodiment, no preprocessing is performed on the data, and directly proceed to the next step.

[0027] S1. Calculation of seismic data attributes According to the input seismic data and the purpose of processing, various statistical domain attributes (including maximum value, minimum value, mean value, median value, skewness, kurtosis, and time difference, etc.), spectral domain attributes (Fourier transform coefficients, wavelet transform parameters, and spectral distance, etc.), and time domain attributes (autocorrelation function, differential mean value, and entropy, etc.) of a single trace can be calculated. Which attributes to calculate depends on the processing purpose and data sensitivity to ensure that the calculated attributes are relevant and sensitive to the data processing process; The property calculated in this embodiment is the time difference. Based on the picked pre-stack direct wave first arrival data, the difference between the theoretical arrival time and the actual arrival time of the direct wave for each trace is calculated, that is, the single-trace time difference, and it is organized according to the geophone domain to form a time difference matrix for each node in the geophone domain.

[0028] If there are multiple single-trace properties with correlation and sensitivity calculated, the principal component analysis method can be used to achieve data dimensionality reduction and reduce the computational complexity; or it can be directly proceeded to the next step without processing. If the single-trace property is a single property, there is no need to perform principal component analysis and directly proceed to the next step. In this embodiment, only the time difference property with correlation and sensitivity is calculated, so there is no need to perform principal component analysis and directly proceed to the next step.

[0029] S2. Seismic data attribute feature extraction The time difference matrix in the geophone domain is used for feature extraction by a deep learning network (a convolutional neural network is used in this embodiment) to generate a feature matrix.

[0030] S3. Seismic data classification Using a clustering analysis algorithm (the K-means algorithm is used in this embodiment), the feature matrix is classified, and then the classification of the nodes is realized. The result is as Figure 3 shown; From Figure 3 it can be seen that after classification, visually and from parameter calculations, the subsea nodes can be divided into 3 categories. Combining with the analysis of actual work experience, it can be known that the characteristics of the first category of nodes are that the XYZ coordinates of the geophone are inaccurate, the characteristics of the second category of nodes are that the XY coordinates of the geophone are inaccurate, and the characteristics of the third category of nodes are that the Z coordinate of the geophone is inaccurate or the water velocity is inaccurate. And the differences between different categories in the classification results are obvious; the nodes within each category are similar or even identical, indicating the effectiveness and accuracy of the classification results.

[0031] S4. Seismic data quality control Applying the classification results of the identified nodes, targeted secondary positioning processing is carried out according to their different characteristics, and the most appropriate positioning parameters are selected to ensure the accuracy and efficiency of positioning (for example, the characteristics of the first category of nodes are that the XYZ coordinates of the geophone are inaccurate, so the data in the XYZ three directions need to be corrected; the characteristics of the second category of nodes are that the XY coordinates are inaccurate, so only the data in the XY directions need to be corrected, and the data in the Z direction does not need to be corrected. This not only greatly reduces the amount of data processing, improves the computational efficiency, but also prevents incorrect processing of data that does not need to be modified, and improves the accuracy of data processing), thereby achieving quality control.

[0032] If the traditional direct wave first arrival positioning method is adopted, directly using the direct wave first arrival information to analyze the overall seismic data through the consistency parameters of space and time, so as to calculate the most likely position of the node, it is necessary to solve an overdetermined linear equation system, which has great calculation difficulty and low accuracy; However, by using the method of this embodiment, the direct wave first arrival data is first deeply processed and classified, and then targeted secondary positioning processing is carried out according to the classification results, which greatly reduces the calculation difficulty and improves the positioning accuracy and efficiency.

[0033] Embodiment 2 This embodiment provides an intelligent quality control device for seismic data based on machine learning, and its structural schematic diagram is as Figure 4 shown, specifically including the following modules connected in sequence: Seismic data attribute calculation module 1: used to calculate single-channel attributes with relevance and sensitivity according to the input seismic data and the purpose of processing, and organize them according to different data domains; Seismic data attribute feature extraction module 2: used to extract features of attributes in different data domains by using a deep learning network to generate a feature matrix; Seismic data classification module 3: using a clustering analysis algorithm to analyze the feature matrix to realize the classification of seismic data; Seismic data quality control module 4: applying the classification results of seismic data to perform targeted data processing to realize the quality control of seismic data.

[0034] Embodiment 3 This embodiment provides a computer device, which includes: a memory, a processor, and a computer program stored on the memory and executable on the processor to implement the above-mentioned intelligent quality control method for seismic data based on machine learning.

[0035] The memory is used to store non-temporary computer-readable instructions. Specifically, the memory may include one or more computer program products, and the computer program products may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory, etc. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0036] The processor may be a central processing unit (CPU) or other forms of processing units with data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions. The processor is used to run the computer-readable instructions stored in the memory.

[0037] Those skilled in the art should be able to understand that, in order to solve the technical problem of how to obtain good user experience effects, well-known structures such as communication buses and interfaces may also be included in this embodiment, and these well-known structures should also be included in the protection scope of this disclosure.

[0038] For the detailed description of this embodiment, reference can be made to the corresponding descriptions in the foregoing embodiments, and details will not be repeated here.

[0039] Embodiment 4 This embodiment provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the intelligent quality control method for seismic data based on machine learning described above is implemented.

[0040] On the computer-readable storage medium, there are non-temporary computer-readable instructions stored. When the non-temporary computer-readable instructions are run by a processor, all or part of the steps of the methods in the foregoing embodiments are executed.

[0041] The above computer-readable storage medium includes but is not limited to: optical storage media (such as CD-ROMs and DVDs), magneto-optical storage media (such as MOs), magnetic storage media (such as magnetic tapes or external hard drives), media with built-in rewritable non-volatile memories (such as memory cards), and media with built-in ROMs (such as ROM cartridges).

Claims

1. A method for intelligent quality control of seismic data based on machine learning, characterized in that: The following steps are involved: S1. Calculate single-channel attributes with correlation and sensitivity according to input seismic data and processing purpose, and organize them according to data domains; S2. Extract features from attributes of different data domains using a deep learning network to generate a feature matrix; S3. Using cluster analysis algorithm, classify the feature matrix to achieve classification of seismic data; S4. Apply the classification results of seismic data to carry out targeted data processing to achieve quality control of seismic data.

2. The intelligent quality control method of seismic data based on machine learning according to claim 1, characterized in that: When the single-channel attribute with correlation and sensitivity calculated in step S1 is multiple attributes, principal component analysis is used to achieve data dimensionality reduction.

3. The intelligent quality control method of seismic data based on machine learning according to claim 1 is characterized in that: It also includes pre-processing the seismic data before step S1.

4. A method for intelligent quality control of seismic data based on machine learning according to any one of claims 1 to 3, characterized in that: The seismic data includes pre-stack data or post-stack data.

5. A method for intelligent quality control of seismic data based on machine learning according to any one of claims 1 to 3, characterized in that: The attributes include statistical domain attributes, spectral domain attributes and time domain attributes.

6. The intelligent quality control method of seismic data based on machine learning according to claim 5, characterized in that: The statistical domain attributes include maximum value, minimum value, mean value, median value, skewness, kurtosis and time difference; the spectral domain attributes include Fourier transform coefficient, wavelet transform parameter and spectrum distance; the time domain attributes include autocorrelation function, differential mean value and entropy.

7. A method for intelligent quality control of seismic data based on machine learning according to any one of claims 1 to 3, characterized in that: The data domain includes a shot domain or a detection point domain.

8. A method for intelligent quality control of seismic data based on machine learning according to any one of claims 1 to 3, characterized in that: The deep learning network includes a convolutional neural network.

9. A method for intelligent quality control of seismic data based on machine learning according to any one of claims 1 to 3, characterized in that: The clustering analysis algorithm includes a K-means algorithm.

10. An intelligent quality control device for seismic data based on machine learning, characterized in that: It consists of the following modules connected in sequence: Seismic data attribute calculation module: used to calculate single-channel attributes with correlation and sensitivity according to the input seismic data and the purpose of processing, and organize them according to different data domains; Seismic data attribute feature extraction module: used to extract features from attributes of different data domains using deep learning networks and generate feature matrices; Seismic data classification module: Use cluster analysis algorithm to classify the feature matrix and realize the classification of seismic data; Seismic data quality control module: applies the classification results of seismic data to carry out targeted data processing to achieve quality control of seismic data.

11. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the intelligent quality control method for seismic data based on machine learning described in any one of claims 1-9 is implemented.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program for executing the intelligent quality control method for seismic data based on machine learning as described in any one of claims 1-9.

Images