A method, system, device, and medium for improving weak signal resolution

Abstract

The application provides a weak signal resolution improvement method, system, device and medium, comprising the following steps: performing frequency band decomposition on a target seismic signal based on deep learning to obtain an effective frequency band signal and a noise frequency band signal; identifying and compensating the effective frequency band signal and the noise frequency band signal under a compressed sensing framework to suppress noise; and reconstructing the effective frequency band signal and the noise frequency band signal processed under the compressed sensing framework to obtain a compensated seismic signal; the application uses the adaptability of deep learning to decompose the deep signal according to the spectral characteristics of the deep signal, extracts and compensates the weak signal of the effective frequency band under the compressed sensing framework, and suppresses the signal of the remaining noise frequency band, thereby accurately suppressing strong noise while protecting the effective signal, especially the deep effective weak signal, and effectively improving the weak signal of the deep structure of the processed profile and enhancing the continuity of the event.

Description

Technical Field

[0001] This invention belongs to the field of petroleum geophysical exploration technology, specifically relating to a method, system, equipment, and medium for improving the resolution of weak signals. Background Technology

[0002] The near-surface conditions in the foreland basin of the western exploration area are exceptionally complex, with extensive penetration of deep salt bodies, narrow bandwidth of seismic data from steep structures and carbonate reservoirs, and poor seismic data quality. Developing identification and adaptive compensation techniques for weak deep signals, and achieving adaptive identification and energy recovery of deep seismic data, is of great research significance for obtaining high-quality mid-to-deep subsurface imaging results.

[0003] Q compensation can enhance the high-frequency components of mid-to-deep layers, effectively improving the quality of seismic data. However, during the compensation process, the amplitude compensation term grows exponentially, especially at high frequencies, making it even more difficult to identify weak deep signals that are already buried by noise.

[0004] Noise suppression is a crucial method for identifying valid signals, and numerous methods have emerged to suppress strong noise, including higher-order statistical methods, stochastic resonance theory methods, singular value decomposition, various sparse transform methods (FFT, wavelet, curvelet, Seislet wave, shearlet wave, etc.), empirical mode decomposition, etc. To improve noise suppression effectiveness, many researchers combine different methods. The recently emerging compressed sensing theory breaks with traditional sampling concepts, being unrestricted by signal bandwidth. Denoising methods based on compressed sensing can utilize the different characteristics of valid signals and noise during compression, and these characteristics can be used to separate valid signals from noise through reconstruction algorithms. However, this method lacks adaptability and cannot effectively separate weak signals from strong interference backgrounds when processing deep seismic data. Summary of the Invention

[0005] In view of the problems existing in the prior art, the present invention provides a method, system, device and medium for improving the resolution of weak signals, which can accurately suppress strong noise while protecting effective signals, especially effective weak signals in deep regions.

[0006] This invention is achieved through the following technical solution:

[0007] A method for improving the resolution of weak signals includes the following steps:

[0008] Based on deep learning, the target seismic signal is decomposed into frequency bands to obtain the effective frequency band signal and the noise frequency band signal.

[0009] Within the framework of compressed sensing, effective frequency band signals and noise frequency band signals are identified and compensated, and noise is suppressed.

[0010] The effective frequency band signal and noise frequency band signal processed under the compressed sensing framework are reconstructed to obtain the compensated seismic signal.

[0011] Preferably, the process of performing frequency band decomposition of seismic signals based on deep learning is as follows:

[0012] Multiple seismic signals are collected and used as training samples to input into a deep learning model, resulting in a trained neural network model.

[0013] The target seismic signal is input into a neural network model and decomposed into different frequency bands.

[0014] Preferably, the process of obtaining the effective frequency band signal and the noise frequency band signal is as follows:

[0015] The effective frequency band signal and the noise frequency band signal are determined based on the autocorrelation analysis method;

[0016] The effective frequency band signal is a frequency band signal whose autocorrelation main lobe width is smaller than the main lobe width of the original noisy data;

[0017] The noise frequency band signal is a frequency band signal whose autocorrelation main lobe width is larger than the main lobe width of the original noisy data.

[0018] Preferably, the process of identifying and compensating for effective frequency band signals and noise frequency band signals within the compressed sensing framework is as follows:

[0019] The quality factor Q is extracted using the spectral ratio method or the frequency shift method, and a compensation operator Λ is set.

[0020] Based on the quality factor Q and using a sparsity enhancement method, the difference between the effective signal and the sensed effective signal is obtained, and noise in the noise band is suppressed.

[0021] min:||Λ(e)||1s.t.||Y-ΦΨe||2≤ε;

[0022] Wherein, Φ is a sparse sampling matrix conforming to the RIP criterion, the sparse sampling matrix including a random Gaussian matrix, a Poisson sampling matrix or a Jitter sampling matrix, Ψ is the sparse transform inverse matrix, including wavelet transform, curvelet transform and / or Seislet transform, e is the sparse transform coefficient, and ε is the noise variance estimate, which is obtained by a greedy algorithm, a basis pursuit algorithm and / or an iterative thresholding algorithm.

[0023] Preferably, the process of reconstructing the effective frequency band signal is as follows:

[0024] The effective signal is reconstructed through the inverse transform of the sparse basis transform:

[0025]

[0026] The compensation operator Λ is used to increase the energy of weak effective signals.

[0027] Preferably, the process of suppressing the effective frequency band signal and the noise frequency band signal within the compressed sensing framework is as follows:

[0028] The difference between the effective signal and the sensed effective signal is obtained through a sparsity enhancement method, and noise in the noise band is suppressed.

[0029] min:||e||1s.t.||Y-ΦΨe||2≤ε;

[0030] Wherein, Φ is a sparse sampling matrix conforming to the RIP criterion, the sparse sampling matrix including a random Gaussian matrix, a Poisson sampling matrix or a Jitter sampling matrix, Ψ is the sparse transform inverse matrix, including wavelet transform, curvelet transform and / or Seislet transform, e is the sparse transform coefficient, and ε is the noise variance estimate, which is obtained by a greedy algorithm, a basis pursuit algorithm and / or an iterative thresholding algorithm.

[0031] Preferably, the process of reconstructing the noise frequency band signal is as follows:

[0032] The effective signal in the noisy frequency band is reconstructed by the inverse transform of the sparse basis transform:

[0033]

[0034] A system for improving weak signal resolution, comprising:

[0035] The preprocessing module is used to perform frequency band decomposition on the target seismic signal based on deep learning to obtain the effective frequency band signal and the noise frequency band signal;

[0036] The compensation module is used to identify and compensate for effective frequency band signals and noise frequency band signals within the compressed sensing framework, and to suppress noise.

[0037] The reconstruction module is used to reconstruct the effective frequency band signal and noise frequency band signal after processing under the compressed sensing framework to obtain the compensated seismic signal.

[0038] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of a method for improving weak signal resolution.

[0039] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of a method for improving weak signal resolution.

[0040] Compared with the prior art, the present invention has the following beneficial technical effects:

[0041] This invention provides a method, system, device, and medium for improving the resolution of weak signals, comprising the following steps: performing frequency band decomposition on a target seismic signal based on deep learning to obtain an effective frequency band signal and a noise frequency band signal; identifying and compensating the effective frequency band signal and the noise frequency band signal under a compressed sensing framework, and suppressing the noise; reconstructing the effective frequency band signal and the noise frequency band signal processed under the compressed sensing framework to obtain a compensated seismic signal; this application uses the adaptivity of deep learning to decompose the deep signal according to the spectral characteristics, extracts and compensates weak signals in the effective frequency band under a compressed sensing framework, and suppresses the signals in the remaining noise frequency bands, thereby accurately suppressing strong noise while protecting the effective signal, especially the effective weak signal in the deep part, and improving the effectiveness and continuity of the phase axis of the deep structure in the processed profile. Attached Figure Description

[0042] Figure 1 This is a flowchart of a method for improving weak signal resolution in an embodiment of the present invention;

[0043] Figure 2 This is a flowchart illustrating a method for improving weak signal resolution in an embodiment of the present invention.

[0044] Figure 3 is a comparison diagram of weak signal compensation in an embodiment of the present invention. Detailed Implementation

[0045] The present invention will be further described in detail below with reference to specific embodiments. These descriptions are for explanation purposes only and are not intended to limit the scope of the invention.

[0046] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0047] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0048] This invention provides a method to improve the resolution of weak signals, such as... Figure 1 and Figure 2 As shown, it includes the following steps:

[0049] Based on deep learning, the target seismic signal is decomposed into frequency bands to obtain the effective frequency band signal and the noise frequency band signal.

[0050] Within the framework of compressed sensing, effective frequency band signals and noise frequency band signals are identified and compensated, and noise is suppressed.

[0051] The effective frequency band signal and noise frequency band signal processed under the compressed sensing framework are reconstructed to obtain the compensated seismic signal.

[0052] Preferably, the process of performing frequency band decomposition of seismic signals based on deep learning is as follows:

[0053] Multiple seismic signals are collected and used as training samples to input into a deep learning model, resulting in a trained neural network model.

[0054] The target seismic signal is input into a neural network model and decomposed into different frequency bands.

[0055] Preferably, the process of obtaining the effective frequency band signal and the noise frequency band signal is as follows:

[0056] The effective frequency band signal and the noise frequency band signal are determined based on the autocorrelation analysis method;

[0057] The effective frequency band signal is a frequency band signal whose autocorrelation main lobe width is smaller than the main lobe width of the original noisy data;

[0058] The noise frequency band signal is a frequency band signal whose autocorrelation main lobe width is larger than the main lobe width of the original noisy data.

[0059] Preferably, the process of identifying and compensating for effective frequency band signals and noise frequency band signals within the compressed sensing framework is as follows:

[0060] The quality factor Q is extracted using the spectral ratio method or the frequency shift method, and a compensation operator Λ is set.

[0061] It should be noted that the process for extracting the quality factor Q using the spectral ratio method provided in this application is as follows:

[0062] Obtaining vibration data: Conducting vibration tests on the vibration system to obtain data such as vibration amplitude or velocity at different frequencies. This can be done using equipment such as vibration test benches and vibration meters.

[0063] Frequency spectrum analysis: Perform frequency spectrum analysis on the acquired vibration data, that is, calculate the vibration amplitude or velocity at each frequency. This can be achieved through mathematical methods such as Fast Fourier Transform (FFT).

[0064] Determine the resonant frequencies: In the frequency spectrum, there will be one or more resonant frequencies. These resonant frequencies are the frequency points at which the vibrating system produces the maximum response to vibrations at a specific frequency. Find these resonant frequencies;

[0065] Calculate the quality factor Q: Use the following formula to calculate the quality factor Q:

[0066] Q = (f1 - f2) / f1 * sqrt(E1 / E2);

[0067] Where f1 and f2 are the low-frequency and high-frequency boundaries of the resonance frequency, and E1 and E2 are the ratios of stored energy to dissipated energy at these boundary frequencies.

[0068] For experimental data, E1 / E2 can be calculated in the following way:

[0069] E1 / E2 = (A1*A2) / (A2-A1)

[0070] Where A1 and A2 are the vibration amplitudes or velocities at frequencies f1 and f2, respectively;

[0071] Calculation result: Substituting f1-f2, f1, E1 and E2 into the above formula, the value of the quality factor Q can be obtained.

[0072] It should be further explained that the quality factor Q is a parameter describing the energy storage and release capability of the vibration system, while the compensation operator is a linear operator introduced into the vibration system to improve the dynamic characteristics and stability of the system. In some cases, the quality factor Q and the compensation operator can influence each other. For example, when there is nonlinear friction in the vibration system, introducing the compensation operator can improve the dynamic characteristics of the system and increase the quality factor Q. This application comprehensively considers the influence of the quality factor Q and the compensation operator in order to better improve the dynamic characteristics and stability of the system.

[0073] Based on the quality factor Q, the difference between the effective signal and the sensed effective signal is obtained through a sparsity enhancement method, and noise in the noise band is suppressed.

[0074] min:||Λ(e)||1s.t.||Y-ΦΨe||2≤ε;

[0075] Wherein, Φ is a sparse sampling matrix conforming to the RIP criterion, the sparse sampling matrix including a random Gaussian matrix, a Poisson sampling matrix or a Jitter sampling matrix, Ψ is the sparse transform inverse matrix, including wavelet transform, curvelet transform and / or Seislet transform, e is the sparse transform coefficient, and ε is the noise variance estimate, which is obtained by a greedy algorithm, a basis pursuit algorithm and / or an iterative thresholding algorithm.

[0076] Preferably, the process of reconstructing the effective frequency band signal is as follows:

[0077] The effective signal is reconstructed through the inverse transform of the sparse basis transform:

[0078]

[0079] The compensation operator Λ is used to increase the energy of weak effective signals.

[0080] Preferably, the process of suppressing the effective frequency band signal and the noise frequency band signal within the compressed sensing framework is as follows:

[0081] The difference between the effective signal and the sensed effective signal is obtained through a sparsity enhancement method, and noise in the noise band is suppressed.

[0082] min:||e||1s.t.||Y-ΦΨe||2≤ε,

[0083] Wherein, Φ is a sparse sampling matrix conforming to the RIP criterion, the sparse sampling matrix including a random Gaussian matrix, a Poisson sampling matrix or a Jitter sampling matrix, Ψ is the sparse transform inverse matrix, including wavelet transform, curvelet transform and / or Seislet transform, e is the sparse transform coefficient, and ε is the noise variance estimate, which is obtained by a greedy algorithm, a basis pursuit algorithm and / or an iterative thresholding algorithm.

[0084] Preferably, the process of reconstructing the noise frequency band signal is as follows:

[0085] The effective signal in the noisy frequency band is reconstructed by the inverse transform of the sparse basis transform:

[0086]

[0087] It should be noted that the random Gaussian matrix is ​​a commonly used matrix. Each row and column of the matrix is ​​generated by a random Gaussian normal distribution. Its characteristics are that the sum of each row and column is zero, and the sum of the squares of the elements in each row or column is proportional to the average of the squares of all elements.

[0088] It should be further explained that the Poisson sampling matrix... A Poisson sampling matrix is ​​a method for achieving efficient sparse representation in signal processing. It leverages the properties of the Poisson distribution to approximately reconstruct high-dimensional signals at a lower dimension. The construction method of a Poisson sampling matrix involves first determining the sparsity of the signal to be reconstructed, i.e., the number of non-zero elements in the signal. Then, the positions of these non-zero elements are randomly selected in the high-dimensional signal. The probability distribution function (PDF) of the Poisson distribution is used in the position selection to obtain positions with a high probability of selecting the number of non-zero elements determined by the sparsity. The advantages of a Poisson sampling matrix include: efficient sparse representation: by utilizing the Poisson distribution, high-dimensional signals can be reconstructed at a lower dimension, thus reducing computational and storage costs; applicability to various signals: the Poisson sampling matrix is ​​applicable to various types of signals, including images, audio, and video; and strong interpretability: the Poisson sampling matrix is ​​constructed based on sparsity and Poisson distribution theory, thus possessing a clear theoretical foundation and interpretability. It should be noted that when using a Poisson sampling matrix for signal reconstruction, additional techniques may be needed to optimize the reconstruction effect, such as least squares methods based on iterative thresholds.

[0089] It should be further explained that Ψ is the inverse matrix of the sparse transformation, which is obtained by wavelet transform, curvelet transform and / or Seislet transform, and the process is as follows:

[0090] This invention provides the general form of a typical wavelet transform function:

[0091]

[0092] Compare the initial part of the wavelet w(t) with the original function f(t) (which is actually an inner product), and calculate the coefficient C. The coefficient C represents the degree of similarity between this part of the function and the wavelet.

[0093] Shift the wavelet to the right by k units to obtain the wavelet w(tk), and repeat step 1. Repeat this step until the function f ends.

[0094] Extend the wavelet w(t) to obtain the wavelet w(t / 2), and repeat steps 1 and 2.

[0095] By continuously expanding the wavelet and repeating the above steps, the sparse transformation inverse matrix Ψ is obtained.

[0096] It should be further explained that ε is the noise variance estimate, which is obtained through a greedy algorithm, a basis pursuit algorithm, and / or an iterative thresholding algorithm. The process is as follows:

[0097] 1. Initialize the estimation results: Set the estimated signal x^ as a zero vector, and denote the initial residual as r = x.

[0098] 2. Select the most relevant atom: Calculate the inner product of the column vectors of D and the residual r, select the atom index i* with the largest inner product value, and update the estimated signal x^, i* = argmax_i

[0099] |<d_i,r> | where d_i is the i-th column vector of dictionary D, x^=x^+

[0100] <d_i*,r> d_i*.

[0101] 3. Update residuals: Calculate the new residuals r = xx^.

[0102] 4. Repeat steps 2 and 3 until the stopping criteria are met (such as the residual threshold or the number of selected atoms reaching a predetermined value).

[0103] 5. Output the final estimated signal x^.

[0104] During the iteration process, OMP progressively selects the atoms with the highest correlation and adds them to the estimated signal, while reducing the energy of the residuals. This allows for the gradual reconstruction of the original sparse signal.

[0105] While the formula derivation is not the core of the OMP algorithm, the update steps in the pseudocode can be derived using the following formula:

[0106] • Calculate the inner product between atoms and residuals:<d_i,r>

[0107] • Calculate the coefficient corresponding to atom i*:<d_i*,r>

[0108] • Update the estimated signal: x^=x^+<d_i*,r> d_i*

[0109] • Update residual: r = r -<d_i*,r> d_i*

[0110] This application utilizes the adaptability of deep learning to decompose deep signals based on their spectral characteristics. It extracts and compensates weak signals in the effective frequency band within the compressed sensing framework, while suppressing signals in the remaining noise frequency bands. This process protects the effective signals, especially the effective weak signals in the deep signal, while accurately suppressing strong noise, as shown in Figure 3. Figure 3(a) shows the signal spectrum of the effective signal in the original data that is obscured by strong interference background. Figure 3(b) shows the signal spectrum of the weak signal that was obscured by strong noise after weak signal extraction and compensation. Figure 3(c) shows the signal spectrum of the weak signal whose energy axis is significantly enhanced after compensation.

[0111] This invention provides a system for improving weak signal resolution, comprising:

[0112] The preprocessing module is used to perform frequency band decomposition on the target seismic signal based on deep learning to obtain the effective frequency band signal and the noise frequency band signal;

[0113] The compensation module is used to identify and compensate for effective frequency band signals and noise frequency band signals within the compressed sensing framework, and to suppress noise.

[0114] The reconstruction module is used to reconstruct the effective frequency band signal and noise frequency band signal after processing under the compressed sensing framework to obtain the compensated seismic signal.

[0115] In another embodiment of the present invention, a computer device is provided, comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in the computer storage medium to achieve a corresponding method flow or corresponding function. The processor described in this embodiment of the present invention can be used in the operation of a method to improve weak signal resolution.

[0116] In another embodiment of the present invention, a storage medium is provided, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the computer device and extended storage media supported by the computer device. The computer-readable storage medium provides storage space that stores the operating system of the terminal. Furthermore, the storage space also stores one or more instructions suitable for loading and execution by a processor, which can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the corresponding steps of the method for improving weak signal resolution in the above embodiments.

[0117] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0118] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations 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 apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations 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.

[0119] 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.

[0120] 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.

[0121] 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 or all of the technical features therein; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for improving the resolution of weak signals, characterized in that, Includes the following steps: Based on deep learning, the target seismic signal is decomposed into frequency bands to obtain the effective frequency band signal and the noise frequency band signal. The process of obtaining the effective frequency band signal and the noise frequency band signal is as follows: The effective frequency band signal and the noise frequency band signal are determined based on the autocorrelation analysis method; The effective frequency band signal is a frequency band signal whose autocorrelation main lobe width is smaller than the main lobe width of the original noisy data; The noise frequency band signal is a frequency band signal whose autocorrelation main lobe width is larger than the main lobe width of the original noisy data; Within the framework of compressed sensing, effective frequency band signals and noise frequency band signals are identified and compensated, and noise is suppressed. The process of identifying and compensating for signals with effective frequency bands and signals with noise frequency bands within the compressed sensing framework is as follows: The quality factor Q is extracted using the spectral ratio method or the frequency shift method, and a compensation operator Λ is set. Based on the quality factor Q and using a sparsity enhancement method, the difference between the effective signal and the sensed effective signal is obtained, and noise in the noise band is suppressed. ； Wherein, Φ is a sparse sampling matrix that conforms to the RIP criterion, and the sparse sampling matrix includes a random Gaussian matrix, a Poisson sampling matrix, or a Jitter sampling matrix; Ψ is the sparse transform inverse matrix, which is obtained by wavelet transform, curvelet transform, and / or Seislet transform; e is the sparse transform coefficient; and ε is the noise variance estimate, which is obtained by a greedy algorithm, a basis pursuit algorithm, and / or an iterative thresholding algorithm. The effective frequency band signal and noise frequency band signal processed under the compressed sensing framework are reconstructed to obtain the compensated seismic signal.

2. The method for improving weak signal resolution according to claim 1, characterized in that, The process of performing frequency band decomposition of seismic signals based on deep learning is as follows: Multiple seismic signals are collected and used as training samples to input into a deep learning model, resulting in a trained neural network model. The target seismic signal is input into a neural network model and decomposed into different frequency bands.

3. The method for improving weak signal resolution according to claim 1, characterized in that, The process of reconstructing the effective frequency band signal is as follows: The effective signal is reconstructed through the inverse transform of the sparse basis transform: ; The compensation operator Λ is used to increase the energy of weak effective signals.

4. A system for improving the resolution of weak signals, characterized in that, A method for improving weak signal resolution according to any one of claims 1-3 includes: The preprocessing module is used to perform frequency band decomposition on the target seismic signal based on deep learning to obtain the effective frequency band signal and the noise frequency band signal; The compensation module is used to identify and compensate for effective frequency band signals and noise frequency band signals within the compressed sensing framework, and to suppress noise. The reconstruction module is used to reconstruct the effective frequency band signal and noise frequency band signal after processing under the compressed sensing framework to obtain the compensated seismic signal.

5. 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 improving weak signal resolution as described in any one of claims 1 to 3.

6. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of a method for improving weak signal resolution as described in any one of claims 1 to 3.

Images