Methods, apparatus, media, and equipment for calculating seismic data absorption attenuation parameters

Abstract

This invention relates to a method, apparatus, medium, and device for calculating absorption attenuation parameters in seismic data. The calculation method includes acquiring seismic data and synthetic seismic data containing labels for absorption attenuation parameters; calculating the Mel spectrum of the seismic data and the synthetic seismic data; inputting the synthetic seismic data into a convolutional neural network model for deep learning training until the loss function value of the convolutional neural network model meets a preset requirement; inputting the Mel spectrum of the seismic data into the convolutional neural network model where the loss function value meets the preset requirement to calculate and obtain the preliminary absorption attenuation parameters of the seismic data; and finally, obtaining the absorption attenuation parameters of the seismic data through Gaussian smoothing. This invention makes the algorithm more stable and more accurate in determining the absorption attenuation parameters by calculating the Mel spectrum and performing Gaussian smoothing on the seismic records. Furthermore, by introducing a deep learning convolutional neural network, it can reduce the error caused by the signal-to-noise ratio of the seismic data.

Description

Technical Field

[0001] This invention relates to the field of petroleum seismic exploration technology, specifically to a method, apparatus, medium, and equipment for calculating seismic data absorption attenuation parameters. Background Technology

[0002] Seismic waves experience energy attenuation during propagation underground, primarily categorized into intrinsic attenuation and non-intrinsic attenuation. Non-intrinsic attenuation is related to the kinematic characteristics of seismic waves, such as geometric diffusion and reflection / transmission losses; intrinsic attenuation is related to the dynamic characteristics of seismic waves, mainly through absorption attenuation. The absorption attenuation by the formation not only weakens the seismic wave energy but also reduces the dominant frequency and bandwidth of the wavelet, affecting the accuracy and resolution of mid-to-deep seismic imaging. However, actual seismic waves are influenced by the underground medium during propagation. When seismic waves propagate through the formation, absorption attenuation causes a reduction in energy, amplitude, and frequency components, with the dominant frequency shifting to lower frequencies as propagation time increases. The absorption attenuation coefficient of the medium is typically used to characterize the absorption attenuation effect on seismic waves during propagation in a model medium; oil and gas-bearing areas often have larger absorption attenuation coefficients.

[0003] When seismic waves propagate through a medium, energy dissipation and velocity dispersion occur. The stronger the absorption and attenuation of seismic waves by the medium, the larger the absorption and attenuation coefficient; conversely, the weaker the absorption and attenuation, the smaller the absorption and attenuation coefficient. Current research has explored various inversion algorithms in the time and frequency domains based on seismic data to obtain the absorption and attenuation coefficient. Generally, time-domain inversion algorithms include amplitude attenuation methods, rise time methods, wavelet simulation methods, and analytic signal methods, while frequency-domain inversion algorithms include spectral ratio methods, matching methods, and spectral simulation methods. Each of these methods has its own applicability and limitations, with frequency-domain inversion algorithms generally offering better applicability and accuracy compared to time-domain inversion algorithms.

[0004] With the rapid development of broadband seismic acquisition and processing technologies, the frequency band of seismic data has been significantly expanded. Currently, the frequency band of broadband seismic data can reach 3–120 Hz, with a bandwidth of up to 5 octaves. Compared with conventional data, broadband seismic data has significant advantages in improving the imaging quality of mid-to-deep seismic data, enhancing the imaging accuracy of complex faults, interpreting sedimentary sequences, and characterizing reservoirs. However, it also presents certain difficulties for estimating the absorption attenuation coefficient of seismic data. For broadband seismic data, the spectral ratio method yields a wide distribution of estimation results, with some estimates deviating significantly from theoretical values. Compared with the spectral ratio method, the centroid frequency shift method yields even wider distributions of estimation results, with many estimates deviating from theoretical values. The peak frequency shift method, on the other hand, shows a significant deviation from theoretical values, reducing the reliability of the absorption attenuation parameter estimation for broadband seismic data. Summary of the Invention

[0005] The purpose of this invention is to provide a method, apparatus, medium, and device for calculating the absorption attenuation parameter of seismic data, so as to solve the problem that the estimation of the absorption attenuation parameter of broadband seismic data deviates from the theoretical value in the existing technology based on various inversion algorithms in the time and frequency domains of seismic data.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A method for calculating seismic data absorption attenuation parameters, comprising:

[0008] Acquire seismic data and synthetic seismic data containing labels for absorption attenuation parameters;

[0009] Calculate the Mel spectrum of seismic data and synthetic seismic data;

[0010] Synthetic seismic data is input into a convolutional neural network model for deep learning training until the loss function value of the convolutional neural network model meets the preset requirements.

[0011] The convolutional neural network model whose loss function value meets the preset requirements is used as the absorption attenuation parameter calculation model. The Mel spectrum of the seismic data is calculated to obtain the preliminary absorption attenuation parameters of the seismic data.

[0012] Gaussian smoothing was applied to the absorption attenuation parameters of the preliminary seismic data to obtain the absorption attenuation parameters of the seismic data.

[0013] Furthermore, it also includes methods for calculating the Mel spectrum of seismic data and synthetic seismic data:

[0014] Seismic data and synthetic seismic data are divided into individual seismic traces into short-time frames. The frequency spectrum of a selected time frame is calculated on the short-time frames using the Fourier transform formula. The Mel spectrum of the selected time frame is obtained by performing a logarithmic calculation on the frequency spectrum. The formula for calculating the Mel spectrum of the selected time frame is as follows:

[0015] ;

[0016] in, To select the Mel spectrum of the time frame, It represents the frequency spectrum.

[0017] Furthermore, it also includes deep learning training methods for inputting synthetic seismic data into convolutional neural network models:

[0018] The Mel spectrum of the synthetic seismic data is used as the raw data for the input layer of the convolutional neural network model. The convolutional and pooling layers of the convolutional neural network model extract features from the raw input data, and the absorption attenuation coefficient of the synthetic seismic data is calculated and output through the fully connected layer of the convolutional neural network model.

[0019] The absorption attenuation parameter labels are used as the learning target data for the output layer of the convolutional neural network model. The absorption attenuation coefficient of the output synthetic seismic data is compared with the absorption attenuation parameter labels in the convolutional neural network model until the absorption attenuation coefficient of the output synthetic seismic data matches the absorption attenuation parameter labels.

[0020] Furthermore, it also includes a method for comparing the absorption attenuation coefficient of the output synthetic seismic data with the absorption attenuation parameter labels: by calculating the loss function value of the absorption attenuation coefficient of the output synthetic seismic data with the absorption attenuation parameter labels, the matching degree between the absorption attenuation coefficient of the output synthetic seismic data and the absorption attenuation parameter labels is determined. The formula for calculating the loss function value is as follows: ,and ;

[0021] in, Represents the loss function. This represents the absorption attenuation coefficient of synthetic seismic data. This indicates the label value of the absorption attenuation parameter. To calculate the number of samples for the loss function, Indicates a point in time.

[0022] Furthermore, it also includes a method for calculating the absorption attenuation parameters of preliminary seismic data: the Mel spectrum of the seismic data whose absorption attenuation parameters are to be calculated is input into a convolutional neural network model whose loss function value meets preset requirements. The convolutional neural network model calculates and outputs the absorption attenuation parameters of the preliminary seismic data. The formula for calculating the absorption attenuation parameters of the preliminary seismic data is as follows:

[0023] ;

[0024] in, For the input data of the convolutional neural network model, These are the absorption attenuation parameters for preliminary seismic data.

[0025] Furthermore, it also includes a Gaussian smoothing method for the absorption attenuation parameters of the preliminary seismic data: The non-stationary absorption attenuation parameters after equalization in the preliminary seismic data are recorded, and the amplitude spectrum at each time point in the time spectrum of the non-stationary absorption attenuation parameters after equalization is smoothed using a Gaussian function. The expression for the Gaussian smoothing function is as follows:

[0026] ;

[0027] in, The absorption attenuation parameters are Gaussian smoothed. The peak value of the absorption attenuation parameter after equalization. The time point for the absorption attenuation parameter after equalization The parameter value, The peak value of the absorption attenuation parameter after equalization. To equalize the time point The absorption attenuation parameter.

[0028] Based on the above-described method for calculating seismic data absorption attenuation parameters, this invention also provides an analysis apparatus, comprising:

[0029] The first processing unit is used to acquire seismic data and synthetic seismic data containing absorption attenuation parameter labels;

[0030] The second processing unit is used to calculate the Mel spectrum of seismic data and synthetic seismic data;

[0031] The third processing unit is used to input synthetic seismic data into a convolutional neural network model for deep learning training until the loss function value of the convolutional neural network model meets the preset requirements.

[0032] The fourth processing unit is used as the absorption attenuation parameter calculation model when the loss function value meets the preset requirements, to calculate the Mel spectrum of the seismic data and obtain the preliminary absorption attenuation parameters of the seismic data.

[0033] The fifth processing unit is used to perform Gaussian smoothing on the absorption attenuation parameters of the preliminary seismic data to obtain the absorption attenuation parameters of the seismic data.

[0034] Based on the above-described method for calculating seismic data absorption attenuation parameters, this invention also provides a computer-readable storage medium storing a computer program thereon, characterized in that the computer program, when executed by a processor, implements the steps of the above-described method for calculating seismic data absorption attenuation parameters.

[0035] Based on the above-described method for calculating seismic data absorption attenuation parameters, this invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the steps of the above-described method for calculating seismic data absorption attenuation parameters.

[0036] The present invention, by adopting the above technical solution, has the following beneficial effects:

[0037] 1. This invention makes the algorithm more stable by performing Mel spectrum calculation and Gaussian smoothing on the seismic records, and can more accurately calculate the absorption attenuation parameters. The algorithm is simple to implement and calculates the absorption attenuation parameters of the seismic data while maintaining the signal-to-noise ratio.

[0038] 2. This invention is based on the introduction of deep learning convolutional neural networks, which can basically eliminate the multi-solution effect caused by compiling absorption attenuation coefficients using synthetic seismic records, ensure the stability of the algorithm, and reduce the error caused by the signal-to-noise ratio of seismic data. It can be widely used in the field of petroleum seismic exploration. Attached Figure Description

[0039] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Throughout the drawings, the same reference numerals denote the same parts. In the drawings:

[0040] Figure 1 This is a schematic block diagram illustrating the steps of a method for calculating seismic data absorption attenuation parameters provided in an embodiment of the present invention;

[0041] Figure 2 These are different synthetic seismic data and their corresponding Mel spectra;

[0042] Figure 3 This is a schematic diagram of deep learning training for a convolutional neural network model;

[0043] Figure 4 It is a curve comparing the absorption attenuation coefficient and the absorption attenuation parameter label of the output synthetic seismic data;

[0044] Figure 5 (a) is a profile of the original seismic data; Figure 5 (b) is a profile of the preliminary seismic data after Gaussian smoothing of the absorption attenuation parameters. Detailed Implementation

[0045] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the invention and to fully convey the scope of the invention to those skilled in the art.

[0046] Existing inversion algorithms based on seismic data in both the time and frequency domains often deviate from theoretical values ​​when estimating absorption attenuation parameters for broadband seismic data. This invention provides a method, apparatus, medium, and device for calculating absorption attenuation parameters of seismic data. The calculation method includes acquiring seismic data and synthetic seismic data containing labels for absorption attenuation parameters; calculating the Mel spectrum of the seismic data and the synthetic seismic data; inputting the synthetic seismic data into a convolutional neural network model for deep learning training until the loss function value of the convolutional neural network model meets a preset requirement; using the convolutional neural network model with the loss function value meeting the preset requirement as the absorption attenuation parameter calculation model; calculating the Mel spectrum of the seismic data to obtain preliminary absorption attenuation parameters; and performing Gaussian smoothing on the preliminary absorption attenuation parameters to obtain the final absorption attenuation parameters of the seismic data. This invention makes the algorithm more stable and more accurate in determining absorption attenuation parameters by calculating the Mel spectrum and performing Gaussian smoothing on the seismic records. Furthermore, by introducing a deep learning convolutional neural network, it can reduce the error caused by the signal-to-noise ratio of the seismic data.

[0047] The present invention will be described in detail below through embodiments.

[0048] Example

[0049] like Figure 1 As shown, this invention provides a method for calculating seismic data absorption attenuation parameters, including:

[0050] Acquire seismic data and synthetic seismic data containing labels for absorption attenuation parameters;

[0051] Calculate the Mel spectrum of seismic data and synthetic seismic data;

[0052] Synthetic seismic data is input into a convolutional neural network model for deep learning training until the loss function value of the convolutional neural network model meets the preset requirements.

[0053] The convolutional neural network model whose loss function value meets the preset requirements is used as the absorption attenuation parameter calculation model. The Mel spectrum of the seismic data is calculated to obtain the preliminary absorption attenuation parameters of the seismic data.

[0054] Gaussian smoothing was applied to the absorption attenuation parameters of the preliminary seismic data to obtain the absorption attenuation parameters of the seismic data.

[0055] like Figure 2 As shown, Figure 2 This section presents different synthetic seismic data and their corresponding Mel spectra.

[0056] Furthermore, it also includes methods for calculating the Mel spectrum of seismic data and synthetic seismic data:

[0057] Seismic data and synthetic seismic data are divided into individual seismic traces into short-time frames. The frequency spectrum of a selected time frame is calculated on the short-time frames using the Fourier transform formula. The Mel spectrum of the selected time frame is obtained by performing a logarithmic calculation on the frequency spectrum. The formula for calculating the Mel spectrum of the selected time frame is as follows:

[0058] ;

[0059] in, To select the Mel spectrum of the time frame, It represents the frequency spectrum.

[0060] As mentioned above, the Fourier transform formula is: ;

[0061] in, For frequency spectrum, For input data, For a specific moment during the earthquake. It is a window function, typically a Hanning window. is the overlap length of the window function.

[0062] Furthermore, it also includes deep learning training methods for inputting synthetic seismic data into convolutional neural network models:

[0063] The Mel spectrum of the synthetic seismic data is used as the raw data for the input layer of the convolutional neural network model. The convolutional and pooling layers of the convolutional neural network model extract features from the raw input data, and the absorption attenuation coefficient of the synthetic seismic data is calculated and output through the fully connected layer of the convolutional neural network model.

[0064] The absorption attenuation parameter labels are used as the learning target data for the output layer of the convolutional neural network model. The absorption attenuation coefficient of the output synthetic seismic data is compared with the absorption attenuation parameter labels in the convolutional neural network model until the absorption attenuation coefficient of the output synthetic seismic data matches the absorption attenuation parameter labels.

[0065] Convolutional neural networks (CNNs) are among the most widely used models in deep learning. A complete CNN typically consists of an input layer, convolutional layers, activation functions, pooling layers, and fully connected layers. The input layer primarily performs preprocessing on the raw data based on the characteristics of the data samples, including operations such as mean removal and normalization.

[0066] Because convolutional kernels are relatively small, downsampling is used to reduce data dimensionality. Pooling layers, simply put, are a form of downsampling; they can significantly reduce data dimensionality. Compared to convolutional layers, pooling layers are more effective at reducing data dimensionality, which not only greatly reduces computational cost but also effectively avoids overfitting.

[0067] In a fully connected layer, all neurons are connected by weights. Fully connected layers are typically located at the end of a convolutional neural network. After the preceding convolutional layers have calculated the features of the data, the next step is classification. Usually, the convolutional network flattens the resulting cuboid at the end into a long vector, which is then fed into the fully connected layer to work with the output layer for classification, thus obtaining the output result.

[0068] like Figure 3 As shown, Figure 3 The left side of the image shows the input Mel spectrum, and the right side shows the output absorption attenuation coefficient.

[0069] Furthermore, it also includes a method for comparing the absorption attenuation coefficient of the output synthetic seismic data with the absorption attenuation parameter labels: by calculating the loss function value of the absorption attenuation coefficient of the output synthetic seismic data with the absorption attenuation parameter labels, the matching degree between the absorption attenuation coefficient of the output synthetic seismic data and the absorption attenuation parameter labels is determined. The formula for calculating the loss function value is as follows: ,and ;in, Represents the loss function. This represents the absorption attenuation coefficient of synthetic seismic data. This indicates the label value of the absorption attenuation parameter. The number of samples is used to calculate the loss function.

[0070] like Figure 4 As shown, Figure 4 This is a curve comparing the absorption attenuation coefficient and the absorption attenuation parameter label of the output synthetic seismic data.

[0071] Furthermore, it also includes a method for calculating the absorption attenuation parameters of preliminary seismic data: the Mel spectrum of the seismic data whose absorption attenuation parameters are to be calculated is input into a convolutional neural network model whose loss function value meets preset requirements. The convolutional neural network model calculates and outputs the absorption attenuation parameters of the preliminary seismic data. The formula for calculating the absorption attenuation parameters of the preliminary seismic data is as follows: ;

[0072] in, For the input data of the convolutional neural network model, These are the absorption attenuation parameters for preliminary seismic data.

[0073] Furthermore, it also includes a Gaussian smoothing method for the absorption attenuation parameters of the preliminary seismic data: The non-stationary absorption attenuation parameters after equalization in the preliminary seismic data are recorded, and the amplitude spectrum at each time point in the time spectrum of the non-stationary absorption attenuation parameters after equalization is smoothed using a Gaussian function. The expression for the Gaussian smoothing function is as follows:

[0074] ;

[0075] in, The absorption attenuation parameters are Gaussian smoothed. The peak value of the absorption attenuation parameter after equalization. The time point for the absorption attenuation parameter after equalization The parameter value, The peak value of the absorption attenuation parameter after equalization. To equalize the time point The absorption attenuation parameter.

[0076] like Figure 5 As shown, Figure 5 (a) is a profile of the original seismic data; Figure 5 (b) is a profile of the absorption attenuation parameters of the preliminary seismic data after Gaussian smoothing.

[0077] This invention provides a method for determining the effective stress coefficient of rocks. By performing Mel spectrum calculation and Gaussian smoothing on seismic records, the algorithm becomes more stable and can more accurately calculate the absorption attenuation parameter. The algorithm is simple to implement and calculates the absorption attenuation parameter of seismic data while maintaining the signal-to-noise ratio. Furthermore, by introducing a deep learning convolutional neural network, the algorithm can largely eliminate the multi-solution effect caused by compiling absorption attenuation coefficients using synthetic seismic records, ensuring the stability of the algorithm and reducing the error caused by the signal-to-noise ratio of seismic data. This method can be widely applied in the field of petroleum seismic exploration.

[0078] Based on the above-described method for calculating seismic data absorption attenuation parameters, this invention also provides an analysis apparatus, comprising:

[0079] The first processing unit is used to acquire seismic data and synthetic seismic data containing absorption attenuation parameter labels;

[0080] The second processing unit is used to calculate the Mel spectrum of seismic data and synthetic seismic data;

[0081] The third processing unit is used to input synthetic seismic data into a convolutional neural network model for deep learning training until the loss function value of the convolutional neural network model meets the preset requirements.

[0082] The fourth processing unit is used as the absorption attenuation parameter calculation model when the loss function value meets the preset requirements, to calculate the Mel spectrum of the seismic data and obtain the preliminary absorption attenuation parameters of the seismic data.

[0083] The fifth processing unit is used to perform Gaussian smoothing on the absorption attenuation parameters of the preliminary seismic data to obtain the absorption attenuation parameters of the seismic data.

[0084] Based on the above-described method for calculating seismic data absorption attenuation parameters, this invention also provides a computer-readable storage medium storing a computer program thereon, characterized in that the computer program, when executed by a processor, implements the steps of the above-described method for calculating seismic data absorption attenuation parameters.

[0085] Based on the above-described method for calculating seismic data absorption attenuation parameters, this invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the steps of the above-described method for calculating seismic data absorption attenuation parameters.

[0086] This invention is described based on flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to specific embodiments. It should be understood that each block of the flowcharts and / or block diagrams, and combinations of blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing device, generate instructions for implementing the flowcharts and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0087] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0088] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0089] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for calculating seismic data absorption attenuation parameters, characterized in that, The calculation method includes: Acquire seismic data and synthetic seismic data containing labels for absorption attenuation parameters; Calculate the Mel spectrum of seismic data and synthetic seismic data; Synthetic seismic data is input into a convolutional neural network model for deep learning training until the loss function value of the convolutional neural network model meets the preset requirements. The convolutional neural network model whose loss function value meets the preset requirements is used as the absorption attenuation parameter calculation model. The Mel spectrum of the seismic data is calculated to obtain the preliminary absorption attenuation parameters of the seismic data. Gaussian smoothing was applied to the absorption attenuation parameters of the preliminary seismic data to obtain the absorption attenuation parameters of the seismic data.

2. The method for calculating seismic data absorption attenuation parameters according to claim 1, characterized in that, It also includes methods for calculating the Mel spectrum of seismic data and synthetic seismic data: Seismic data and synthetic seismic data are divided into individual seismic traces into short-time frames. The frequency spectrum of a selected time frame is calculated on the short-time frames using the Fourier transform formula. The Mel spectrum of the selected time frame is obtained by performing a logarithmic calculation on the frequency spectrum. The formula for calculating the Mel spectrum of the selected time frame is as follows: ； in, To select the Mel spectrum of the time frame, It represents the frequency spectrum.

3. The method for calculating seismic data absorption attenuation parameters according to claim 2, characterized in that, It also includes deep learning training methods for inputting synthetic seismic data into convolutional neural network models: The Mel spectrum of the synthetic seismic data is used as the raw data for the input layer of the convolutional neural network model. The convolutional and pooling layers of the convolutional neural network model extract features from the raw input data, and the absorption attenuation coefficient of the synthetic seismic data is calculated and output through the fully connected layer of the convolutional neural network model. The absorption attenuation parameter labels are used as the learning target data for the output layer of the convolutional neural network model. The absorption attenuation coefficient of the output synthetic seismic data is compared with the absorption attenuation parameter labels in the convolutional neural network model until the absorption attenuation coefficient of the output synthetic seismic data matches the absorption attenuation parameter labels.

4. The method for calculating seismic data absorption attenuation parameters according to claim 3, characterized in that, It also includes a method for comparing the absorption attenuation coefficient of the output synthetic seismic data with the absorption attenuation parameter labels: by calculating the loss function value of the absorption attenuation coefficient of the output synthetic seismic data with the absorption attenuation parameter labels, the matching degree between the absorption attenuation coefficient of the output synthetic seismic data and the absorption attenuation parameter labels is determined. The formula for calculating the loss function value is as follows: ,and ; in, Represents the loss function. This represents the absorption attenuation coefficient of synthetic seismic data. This indicates the label value of the absorption attenuation parameter. To calculate the number of samples for the loss function, Indicates a point in time.

5. The method for calculating seismic data absorption attenuation parameters according to claim 4, characterized in that, It also includes a method for calculating the absorption attenuation parameters of preliminary seismic data: The Mel spectrum of the seismic data whose absorption attenuation parameters are to be calculated is input into a convolutional neural network model whose loss function value meets preset requirements. The convolutional neural network model calculates and outputs the absorption attenuation parameters of the preliminary seismic data. The formula for calculating the absorption attenuation parameters of the preliminary seismic data is as follows: ; in, For the input data of the convolutional neural network model, These are the absorption attenuation parameters for preliminary seismic data.

6. The method for calculating seismic data absorption attenuation parameters according to claim 5, characterized in that, It also includes a Gaussian smoothing method for the absorption attenuation parameters of the preliminary seismic data: Record the non-stationary absorption attenuation parameters after equalization in the preliminary seismic data, and smooth the amplitude spectrum of each time point in the time spectrum of the non-stationary absorption attenuation parameters after equalization using a Gaussian function. The function expression for Gaussian smoothing is: ； in, The absorption attenuation parameters are Gaussian smoothed. The peak value of the absorption attenuation parameter after equalization. The time point for the absorption attenuation parameter after equalization The parameter value, The peak value of the absorption attenuation parameter after equalization. For the time point before equilibrium The absorption attenuation parameter.

7. An analytical apparatus, characterized in that, The analytical apparatus includes: The first processing unit is used to acquire seismic data and synthetic seismic data containing absorption attenuation parameter labels; The second processing unit is used to calculate the Mel spectrum of seismic data and synthetic seismic data; The third processing unit is used to input synthetic seismic data into a convolutional neural network model for deep learning training until the loss function value of the convolutional neural network model meets the preset requirements. The fourth processing unit is used as the absorption attenuation parameter calculation model when the loss function value meets the preset requirements, to calculate the Mel spectrum of the seismic data and obtain the preliminary absorption attenuation parameters of the seismic data. The fifth processing unit is used to perform Gaussian smoothing on the absorption attenuation parameters of the preliminary seismic data to obtain the absorption attenuation parameters of the seismic data.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for calculating the seismic data absorption attenuation parameter according to any one of claims 1-6.

9. 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, it implements the steps of the method for calculating the seismic data absorption attenuation parameter according to any one of claims 1-6.

Images