First arrival refraction area linear interference suppression method, product, medium and equipment

Abstract

The invention discloses a first arrival refraction area linear interference suppression method, a product, a medium and equipment, and belongs to the technical field of seismic data processing. The method comprises the following steps: designing different parameters aiming at inconsistent linear characteristics of different frequency band ranges; the multi-scale and multi-directional representation capability of curvelet transform is utilized to decompose seismic data into subsets with different scales and directions, and then accurate wave field separation is realized through Radon transform in each curvelet domain, so that refraction interference with relatively strong energy in a first arrival refraction region is eliminated, and effective reflection signals in a large offset range are recovered. The method can accurately suppress first-arrival refraction interference, recover effective signals, and improve seismic imaging quality.

Description

Technical Field

[0001] This invention belongs to the field of seismic data processing technology, specifically relating to a method, product, medium, and equipment for suppressing linear interference in the first arrival refraction zone. Background Technology

[0002] In marine towed seismic acquisition, water depth significantly affects the linear travel-time curve of first-arrival refracted waves. According to Snell's law, the horizontal distance required for a seismic wave to strike the seabed at a critical angle and generate a refracted wave is very short. Therefore, a receiver very close to the source can receive refracted waves from a high-velocity basement. Due to the shallow water, the travel time difference between the refracted wave path and the direct wave path is small. If the formation velocity is much greater than the water velocity, the refracted wave arrives earlier than the direct wave. Therefore, any reflected wave from the same interface or a shallower interface will necessarily have a later travel time than the refracted wave. On the travel-time curve, the travel time of the refracted wave shows a good linear relationship with the shot-receiver distance. The first-arrival zone, dominated by refracted waves, is quickly replaced by refracted waves, forming a clear and stable linear region. Furthermore, since refracted waves propagate along the interface above the high-velocity layer, they usually accumulate a large amount of energy, especially refracted waves generated at the top of the high-velocity basement, which have very strong energy. In contrast, the energy of reflected waves decreases with the square of the propagation distance, and some energy is transmitted at the interface, making them relatively weak.

[0003] Figure 1 For a single-shot forward model, it can be seen that within a cable length range of 7,000 to 15,000 meters, the first arrival refraction interference range accounts for almost 50% of the single-shot range, and the range of the gun ranging below 4 seconds is almost submerged in the linear refraction interference.

[0004] Seismic data acquired via towed cable in shallow water exhibits a triangular first-arrival refraction zone with strong interference wave energy and a relatively good linear pattern. This type of linear interference appears as a low-velocity characteristic on the original single-shot data. Typically, techniques such as Radon transform delinearization, FK delinearization, and TX domain tilt filtering are used to suppress linear noise within the triangular zone where first-arrival refraction interference develops, based on the apparent velocity difference between linear interference and the effective signal. However, due to the complexity of the interference waves in the first-arrival refraction zone, these techniques often result in incomplete suppression or loss of effective signal during the suppression process.

[0005] Therefore, a new method for suppressing linear interference in the first-arrival refraction region is needed. Summary of the Invention

[0006] The present invention aims to at least partially solve one of the technical problems in the aforementioned related technologies.

[0007] Therefore, the purpose of this invention is to provide a method, product, medium, and device for suppressing linear interference in the first arrival refraction zone, which can accurately suppress first arrival refraction interference, restore effective signals, and improve the quality of seismic imaging.

[0008] To solve the above-mentioned technical problems, the present invention is implemented as follows: This invention provides a method for suppressing linear interference in the first-arrival refraction region, the method comprising: Different parameters are designed to address the inconsistent linear characteristics across different frequency bands. By utilizing the multi-scale and multi-directional characterization capabilities of curve wave transform, seismic data is decomposed into subsets of different scales and directions. Then, precise wavefield separation is achieved within each curve wave domain using Radon transform to eliminate strong refraction interference in the first arrival refraction zone and recover the effective reflection signal within the gun-receiver range.

[0009] In addition, the linear interference suppression method for the first-arrival refraction region according to the present invention may also have the following additional technical features: In some embodiments, the steps of the method include: S1. Preprocessing and fine-grained frequency band division of seismic data; S2. Differentiated parameter design for different frequency bands; S3. Perform a curvelet transform on the seismic data according to the design parameters, convert the tx domain data into the curvelet domain and decompose it into subsets of different scales and directions; S4. Based on the design parameters, in each curve wave domain, the corresponding frequency band of the Radon transform parameters are applied sequentially to map the linear interference in the curve wave domain to the focusing energy point in the Radon domain. The linear Radon transform is realized by using a discretized integral form, and the ill-posedness of the discrete Radon transform is solved by least squares regularization. Finally, the Radon domain data corresponding to each subset of the curve wave domain is obtained, and accurate wave field separation is achieved through Radon transform. S5. Using the conjugate transpose operator of the Radon transform, the Radon domain data is transformed back to the curve wave coefficient domain, and the curve wave coefficient matrix of each scale-direction is recovered. At this time, only the effective wave coefficients and a small amount of residual noise coefficients are retained in the matrix. S6. Perform inverse curvelet transform on the recovered curvelet coefficient matrix.

[0010] In some implementations, Ladon domain adaptive threshold filtering is performed between steps S4 and S5: For each data block in the Ladon domain, first calculate its standard deviation and determine the threshold according to the frequency band. Then, use the corresponding type of threshold filtering to process the data point by point to remove the focused energy of the interference wave. After processing, perform energy statistics on the data. If the interference wave energy removal rate is less than the judgment threshold, then readjust the threshold to ensure the interference suppression effect.

[0011] In some implementations, step S3 includes: 2D FFT transformation yields frequency-wavenumber domain data; Non-uniform resampling is performed according to the curvature scale parameter J; Weighted scaling with a Gaussian window function can achieve scale or orientation localization. Wrapping around the origin to locally concentrate energy; The 2D IFFT transform yields the curvelet coefficient matrix; The final output is a curve coefficient matrix of different scales and directions corresponding to low, medium and high frequency bands.

[0012] In some implementations, step S2 includes: Curved wave transform parameters: The second-generation Wrapping fast algorithm is used to design the number of scales. J , direction number L Window function bandwidth coefficient α Low-frequency bands have coarse scale and low directional resolution, while high-frequency bands have fine scale and high directional resolution, in order to adapt to the degree of fragmentation of the linear features of the interference. Radon transform parameters: A linear Radon transform is used, with the core design focusing on ray parameters. p Regularization parameters λ ray parameters p The regularization parameter λ is negatively correlated with the apparent velocity and positively correlated with the interference energy, in order to address the ill-posedness of the Radon transform and improve the focusing effect.

[0013] In some implementations, step S2 further includes: Threshold filtering parameters: Based on the energy difference between interference and effective waves, three types of thresholds are designed: hard threshold, semi-soft threshold, and soft threshold. Different threshold sizes and coefficients are used to balance the thoroughness of interference removal and the amplitude preservation of effective signals.

[0014] In some implementations, step S6 includes: First, a 2D FFT transformation is performed on the denoised curve coefficient matrix. The original position in the frequency domain is restored by inverse wrapping operation, and scale or direction inverse localization is completed by conjugate multiplication with the window function. Finally, all scale-direction data blocks are merged and transformed back to the tx domain by 2DIFFT transformation to obtain single-shot seismic data with first-arrival refraction interference removed.

[0015] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the linear interference suppression method for the initial refraction region as described above.

[0016] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the linear interference suppression method for the first arrival refraction region as described above.

[0017] The present invention also provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the linear interference suppression method for the first arrival refraction region as described above.

[0018] Compared with the prior art, the present invention has at least the following beneficial effects: The linear interference suppression method in the first-arrival refraction region provided in this embodiment of the invention can solve the problem that conventional linear interference suppression techniques are prone to incomplete interference suppression or loss of effective signals, and overcome the drawbacks of poor energy focusing and spatial aliasing in τ-p transform. This invention uses high-resolution Radon transform in the curve domain, which avoids the defects of conventional techniques in principle and has stronger technical applicability. In this embodiment of the invention, the provided method for suppressing linear interference in the first-arrival refraction zone utilizes the superior multi-scale and multi-directional characterization capabilities of curve wave transform to decompose seismic data into subsets of different scales and directions. Then, within each curve wave domain, a high-precision Radon transform is used to achieve accurate separation of the wavefield, which can specifically eliminate linear interference with strong energy in the first-arrival refraction zone and avoid misprocessing of effective signals. In this embodiment of the invention, the provided method for suppressing linear interference in the first-arrival refraction region removes linear interference more thoroughly after processing, while providing better protection for information at large incident angles. It effectively reduces the loss of effective signals and significantly improves fidelity compared to conventional techniques. In this embodiment of the invention, the provided method for suppressing linear interference in the first-arrival refraction zone can effectively recover the effective reflection signals of the gun-receiver distance and mid-to-long-range paths, providing more reliable data support for steep-angle tomography and mid-to-deep imaging, and significantly improving the overall imaging effect.

[0019] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0020] Figure 1 This is a publicly available forward modeling diagram of a single shot; Figure 2 This is a full-frequency single-gun diagram disclosed in one embodiment of the present invention; Figure 3 A single-shot comparison diagram of different refraction interference suppression methods disclosed in an embodiment of the present invention; Figure 4 This is a comparison diagram of a single gun before and after different refraction interference suppression, as disclosed in an embodiment of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings and specific examples and application scenarios.

[0023] This invention first conducts frequency-division analysis on the original data, investigates and measures the linear velocity range of shallow refraction, and determines the distribution area and velocity range of first-arrival refraction interference in different frequency bands. Specifically, it designs different parameters to address the inconsistent linear characteristics across different frequency bands. The investigation revealed that first-arrival refraction interference does not develop within the triangular region in different frequency bands, such as... Figure 2 As shown, on a full-frequency single shot, the first arrival refraction interference develops within the triangular region, but on single-shot records below 10Hz, the range of refraction linear interference development is wider (blue line area).

[0024] This invention employs a method for suppressing refracted interference waves in high-resolution Radon transform seismic data in the curve wave domain. This method effectively avoids problems such as poor energy focusing and spatial aliasing associated with the τ-p transform. By utilizing the superior multi-scale and multi-directional characterization capabilities of the curve wave domain Radon transform, the seismic data is decomposed into subsets of different scales and directions. Within each curve wave domain, precise wavefield separation is achieved through high-precision Radon transform, effectively eliminating strong refracted interference in the first-arrival refraction zone and effectively recovering the effective reflection signal within the gun-receiver range. Figure 3 By comparing single-shot (after dynamic correction) before and after suppression of refraction interference using different methods, it can be seen that the present invention removes refraction linear interference more thoroughly, protects information at large incident angles better, and has higher technical fidelity. Figure 4 The comparison between this technique and two conventional techniques after single-shot processing shows that the effective signal at medium and long ranges is well recovered, which is of great significance for improving steep-angle tomography and medium-deep imaging.

[0025] (a) Design and implementation of differentiated parameters to address the inconsistent linear characteristics of different frequency bands: Linear interference (surface waves, direct waves, and multiple refracted waves) and effective reflected waves in seismic data exhibit significant heterogeneity in their linear characteristics across different frequency bands: low-frequency interference waves show extremely strong linearity and concentrated energy; mid-frequency interference waves exhibit differentiated linear characteristics; and high-frequency interference waves show degraded linearity while highlighting details in the effective waves. The core of this work is to establish a closed-loop system of "frequency band division - feature quantization - parameter mapping - optimization verification," configuring frequency band-specific parameters for curve transform and Radon transform to achieve accurate and effective interference suppression.

[0026] 1) Preprocessing and fine-grained frequency band division: Preprocessing steps: The raw single-shot seismic data is subjected to channel equalization, DC component removal, and spectral correction to eliminate gain bias and spectral distortion of the acquisition equipment; the global spectral characteristics of the data are obtained through time-frequency analysis (such as S-transform) to determine the dominant frequency range of the work area (usually 5-80Hz in the Loess Plateau region).

[0027] Frequency band division principle: The seismic data is divided into three core frequency bands based on both geological significance and linear characteristics. The division threshold is based on the spectral statistics and interference wave identification results of typical single shots in the work area. Low frequency band (5-20Hz): The core interference is surface wave, and the energy ratio can reach more than 60% of the original data. The phase axis is strictly linear, and the apparent velocity is stable at 300-600m / s. Mid-frequency band (20-50Hz): Direct waves, multiple refracted waves and effective reflected waves in the middle and shallow layers are mixed. The linearity of the interference waves is moderate and the apparent velocity range is large (600-2500m / s). High frequency band (50-80Hz): The energy of the interference wave decays rapidly, the linear characteristics break down, the thin-layer information and structural details of the effective reflected wave become dominant, and the proportion of random noise increases significantly.

[0028] Implementation method: The frequency division is completed by using a bandpass filter bank. The filter is a zero-phase Chebyshev type II filter with a passband ripple of ≤0.5dB and a stopband attenuation of ≥40dB for each frequency band. This avoids the introduction of phase distortion during the frequency division process and ensures the authenticity of the linear characteristics of the data in each frequency band.

[0029] Quantitative characterization of linear characteristics in each frequency band: For each frequency band, the characteristics are quantified from three dimensions: linearity, energy concentration, and apparent velocity distribution, providing a quantitative basis for parameter design. Specifically: Linearity Measurement: The in-phase axis fitting error is used as the core indicator. Linear fitting is performed on the in-phase axis in the tx domain of single-shot data for each frequency band, and the root mean square (RMS) of the fitting residuals is calculated. Low frequency band: fitting residual RMS < 0.02 ms, linearity ≥ 98%; Mid frequency band: fitting residual RMS between 0.02 and 0.08 ms, linearity 70%-98%; High frequency band: fitting residual RMS > 0.08 ms, linearity < 70%.

[0030] Energy concentration analysis: characterized by frequency band energy proportion and spatial coherence. Low-frequency interference wave energy is highly coherent in space, with a coherence coefficient ≥ 0.9; mid-frequency band coherence coefficient is 0.5-0.9; high-frequency band coherence coefficient is < 0.5, and the energy is diffuse.

[0031] Apparent velocity distribution statistics: Velocity spectrum analysis was used to obtain the apparent velocity range of interference waves in each frequency band, which is used for the ray parameters of the Radon transform. p The settings provide boundary conditions.

[0032] 2) Design of frequency band-specific parameters for curve transform The curve transform employs a second-generation fast wrapping algorithm, whose parameters directly determine the ability to capture linear features in different frequency bands. The core parameter design logic is as follows: Scale number J: Positively correlated with frequency band. The low-frequency band corresponds to the coarse-scale characteristics of seismic data, so a small scale number (J=3) is set to focus on strong energy linear interference; the mid-frequency band increases the scale number (J=4) to take into account the separation of interference and effective waves; the high-frequency band uses a multi-scale number (J=5) to capture broken linear interference and effective wave details through fine scale.

[0033] Directional number L: Negatively correlated with the linearity of the interference wave. Low-frequency interference waves have extremely high linearity, so high directional resolution is not required, and L=8 is set; mid-frequency interference waves exhibit differentiated linear characteristics, so L=16 is set; high-frequency interference waves exhibit linear fragmentation, requiring high directional resolution to distinguish between effective waves and interference, so L=32 is set.

[0034] Window function bandwidth coefficient α: controls the width of the window function in the frequency-wavenumber domain and is inversely proportional to the bandwidth. In the low-frequency band, where the bandwidth is narrow, α is set to 0.8 to expand the window function coverage; in the high-frequency band, where the bandwidth is wide, α is set to 0.4 to improve frequency resolution and avoid aliasing of different frequency components.

[0035] 3) Design of band-specific parameters for Radon transform: A linear Radon transform was employed (adapted to the time-distance curve characteristics of linear disturbances in the Loess Plateau region), with core parameters including ray parameters. p The regularization parameter λ is designed strictly according to the apparent velocity and energy characteristics of the frequency band. Ray parameters p Defined as p =1 / v ( v The apparent velocity (LAV) and its range, along with the sampling interval, directly determine the focusing effect in the Radon domain. Low-frequency surface wave apparent velocities are low (300-600 m / s). p The range is [0.0016, 0.0033] s / m; the apparent velocity of the direct wave in the mid-frequency band is moderate (600-2500m / s). p The range is [0.0004, 0.0016] s / m; the high-frequency interference wave has a high apparent velocity. p The range is narrowed to [0.0001, 0.0004] s / m. Sampling interval Δ pProportional to frequency, a smaller spacing (0.00002 s / m) is used in the high-frequency band to ensure high resolution in the Radon domain.

[0036] The regularization parameter λ is used to address the ill-posedness of the Radon transform and is positively correlated with the energy of the interference wave. In the low-frequency band, where interference energy is strong, λ is set to 0.01 to enhance the regularization effect and avoid overfitting; in the high-frequency band, where interference energy is weak, λ is set to 0.001 to preserve the weak signal of the effective wave.

[0037] 4) Design of frequency band-specific parameters for threshold filtering Threshold filtering is crucial for Radon domain interference removal. Different threshold types and sizes are used based on the energy difference between the interfering and active waves in each frequency band. Low frequency band: Interference waves in the Radon domain appear as focal points of extremely strong energy, which are significantly different from the energy of the effective wave. A hard threshold is used (data greater than the threshold is set to zero, and data less than the threshold is retained), with a threshold τ=3σ, to ensure that strong interference is completely eliminated. Mid-frequency band: Interference and effective wave energy are mixed. A semi-soft threshold is used, and a transition zone is set at the threshold τ=2σ to balance interference suppression and effective wave amplitude preservation. High frequency band: The effective wave energy is weak. A soft threshold (shrunk the data proportionally) is used. The threshold τ=1.2σ and the shrinkage coefficient γ=0.8 to preserve the details of the effective wave to the maximum extent while suppressing random noise.

[0038] After the parameter design is completed, its effectiveness must be ensured through a process of typical single-shot testing, iterative optimization, and batch verification in the work area. Typical single shot selection: Select 3-5 representative single shots (including strong surface waves, direct waves and complex structural reflections) from different locations in the work area to cover all geological units; Parameter Iteration Optimization: Perform frequency division processing and curve domain Radon transform on a typical single shot, calculate the signal-to-noise ratio (SNR) and effective wave amplitude retention rate of the denoised data. If the SNR improvement is <15dB or the amplitude retention rate is <80%, adjust the parameters of the corresponding frequency band (such as fine-tuning the ray parameter range and threshold size) until the index is met. Batch verification in the work area: The optimized parameters are applied to the entire work area, and the superimposed profiles before and after noise reduction are compared to verify the consistency of the interference suppression effect and ensure that the weak reflection structure of the Loess Plateau is not destroyed.

[0039] (II) Full-process implementation of interference suppression using the curvilinear domain Radon transform The core logic of the curve wave domain Radon transform interference suppression technology is "curve wave domain separation + Radon domain focusing + threshold domain removal": Utilizing the multi-scale and multi-directional characteristics of the curve wave transform, linear interference in seismic data maintains its linear characteristics in the curve wave domain, while the effective waves are dispersed; then, the linear interference in the curve wave domain is focused into energy points / energy clusters in the Radon domain through Radon transform; finally, threshold filtering removes the interference, and inverse transform restores high signal-to-noise ratio seismic data. Combining patented technical solutions and engineering implementation, the entire process consists of 5 core steps and includes several key technical details.

[0040] Curved wave transform, as a multi-scale analysis method with direction sensitivity and scale sparsity, can map linear interference to continuous coefficients with strong directionality in the curved wave domain, while the effective wave is decomposed into dispersed, low-energy coefficients. Radon transform (especially linear Radon transform) can focus the linear in-phase axis in the tx domain into an energy point in the Radon domain (τ-p domain), achieving complete separation of interference waves and effective waves.

[0041] Step 1: Preprocessing of raw seismic data and curvelet transform The curve forward transform employs the second-generation Wrapping fast algorithm, which is based on the two-dimensional Fourier transform (2DFFT), has high computational efficiency, and is suitable for large-scale seismic data processing. The specific implementation consists of five sub-steps: 2D FFT Transformation: For preprocessed single-shot seismic data ( t For time, x A two-dimensional Fourier transform is performed on the coordinates of the receiver point to obtain frequency-wavenumber domain data. ,in ω Angular frequency, k Let be the wave number. Mathematically, it is expressed as: .

[0042] Frequency-wavenumber domain resampling: based on the scaling parameters of the curve transform. J ,right Non-uniform resampling is performed to divide the frequency domain into different scale regions, with each scale corresponding to a frequency range, consistent with the previous frequency band division.

[0043] Window function weighting: Weighting the resampled data with a curvelet window function. Multiplication achieves localization of scale and direction. The window function is composed of a scale window. With the direction window Composition. Mathematically expressed as: ,in , , j For scale indexing, l For direction index.

[0044] Engineering implementation: The window function adopts a Gaussian window, and its bandwidth coefficient α is set according to the frequency band parameters to ensure accurate scale and direction coverage of different frequency bands.

[0045] Localization around the origin: The data after weighting by the window function is folded around the origin of the frequency domain, concentrating the dispersed energy in the central region, reducing data redundancy and improving computational efficiency.

[0046] 2D IFFT Transform: Perform a two-dimensional inverse Fourier transform on each scale-direction data block after wrapping to obtain the curvelet coefficient matrix. C ( j , l , i ),in i For spatial location index.

[0047] Output results: Curvature coefficient matrices at different scales and in different directions are obtained. Each matrix corresponds to the linear characteristics of a frequency band, with the low-frequency band corresponding to the coarse scale. j =1-3), low direction number ( l A matrix of (1-8) corresponds to a finer scale in the high-frequency band. j =3-5), high direction number ( l A matrix of 17-32.

[0048] Step 2: Scale / Direction Radon Transform For each curve wave coefficient matrix obtained in step 1, a linear Radon transform is performed according to the designed frequency band-specific Radon transform parameters to map the linear interference in the curve wave domain to the focused energy point in the Radon domain.

[0049] 2.1 Transformation Object: Pair the curve wave coefficient matrix for each scale-direction one by one. C ( j , l , i The parameters are processed to ensure that they are accurately matched with the frequency bands corresponding to the matrix.

[0050] 2.2 Mathematical Implementation of Linear Radon Transform: Using a discretized integral form, the curvelet coefficients in the tx domain are mapped to τ- p Ladon coefficient of the domain U (τ, p ), where τ is the intercept time, p Ray parameters: Continuous form:

[0051] Discrete form: ,in m=1,2,..., , n =1,2,..., , k =1,2,..., , This represents the distance between detector points.

[0052] High-resolution solution: Since the Discrete Radon Transform is an ill-posed problem, least squares regularization is used for the solution. The objective function is: ; Solution results: ,in L For the Radon transform operator, L H It is its conjugate transpose operator (inverse transformation operator). I It is the identity matrix. λ This is the regularization parameter (set according to the frequency band parameter).

[0053] Output: The Radon domain data U corresponding to each curve coefficient matrix is ​​obtained. j,l (τ, p In the Radon domain, low-frequency interference waves appear as focal points of extremely strong energy, while high-frequency interference waves appear as dispersed energy clusters, and effective waves exhibit a diffuse distribution of weak energy.

[0054] Step 3: Radon domain adaptive threshold filtering Threshold filtering is the core of interference suppression. According to the designed frequency band-specific threshold parameters, each Radon domain data block is filtered to remove the focused energy of the interference wave and retain the weak signal of the effective wave.

[0055] Threshold calculation: First, calculate the standard deviation σ of each Radon domain data block, and then determine the threshold τ according to the frequency band type (e.g., τ=3σ for low frequency band).

[0056] Filtering implementation: Process Radon domain data point by point according to the threshold type: Hard threshold (low frequency band): ; Semi-soft threshold (mid-frequency band): ,in k Transition coefficient; Soft threshold (high frequency band): ,in γ This is the shrinkage coefficient.

[0057] Validity test: For the filtered Radon domain data Energy statistics are performed. If the interference wave energy rejection rate is less than 85%, the threshold value is readjusted until the requirements are met.

[0058] Step 4: Radon Inverse Transformation Filtered Radon domain data The transformation back to the curvelet coefficient domain restores the curvelet coefficient matrix for each scale-direction. This is specifically achieved using the conjugate transpose operator of the Radon transform.

[0059] Mathematical expression: ; Discrete implementation: ,in, Set the sampling interval for ray parameters (based on frequency band parameters).

[0060] Output result: The denoised curve coefficient matrix is ​​obtained. At this point, the matrix retains only the coefficients of the effective wave and a small amount of residual noise coefficients.

[0061] Step 5: Inverse Curve Wave Transform The inverse transform of the denoised curvelet coefficient matrix is ​​performed to obtain the denoised seismic data in the time-space domain, completing the entire interference suppression process. The inverse curvelet transform is the inverse process of the forward curvelet transform, and the specific steps are as follows: For each denoised curve coefficient matrix Perform a two-dimensional Fourier transform to obtain frequency-wavenumber domain data; Perform inverse wrapping on the data to restore its original position in the frequency domain; With window function The conjugate multiplication completes the inverse localization of scale and direction; All scale-direction data blocks are merged, and then a two-dimensional inverse Fourier transform is performed to obtain denoised single-shot seismic data. .

[0062] The above process was applied to all single-shot data in the work area, and then superimposed to obtain a high signal-to-noise ratio seismic superimposed profile.

[0063] For the parts of this invention not described in detail, please refer to the prior art or the art known to those skilled in the art. This embodiment does not limit these aspects and will not describe them in detail here.

[0064] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of the present invention.

Claims

1. A method for suppressing linear interference in the first-arrival refraction region, characterized in that, The method includes: Different parameters are designed to address the inconsistent linear characteristics across different frequency bands. By utilizing the multi-scale and multi-directional characterization capabilities of curve wave transform, seismic data is decomposed into subsets of different scales and directions. Then, precise wavefield separation is achieved within each curve wave domain using Radon transform to eliminate strong refraction interference in the first arrival refraction zone and recover the effective reflection signal within the gun-receiver range.

2. The method for suppressing linear interference in the first-arrival refraction region according to claim 1, characterized in that, The steps of the method include: S1. Preprocessing and fine-grained frequency band division of seismic data; S2. Differentiated parameter design for different frequency bands; S3. Perform a curvelet transform on the seismic data according to the design parameters, convert the tx domain data into the curvelet domain and decompose it into subsets of different scales and directions; S4. Based on the design parameters, in each curve wave domain, the corresponding frequency band of the Radon transform parameters are applied sequentially to map the linear interference in the curve wave domain to the focusing energy point in the Radon domain. The linear Radon transform is realized by using a discretized integral form, and the ill-posedness of the discrete Radon transform is solved by least squares regularization. Finally, the Radon domain data corresponding to each subset of the curve wave domain is obtained, and accurate wave field separation is achieved through Radon transform. S5. Using the conjugate transpose operator of the Radon transform, the Radon domain data is transformed back to the curve wave coefficient domain, and the curve wave coefficient matrix of each scale-direction is recovered. At this time, only the effective wave coefficients and a small amount of residual noise coefficients are retained in the matrix. S6. Perform inverse curvelet transform on the recovered curvelet coefficient matrix.

3. The method for suppressing linear interference in the first-arrival refraction region according to claim 2, characterized in that, Ladon domain adaptive threshold filtering is performed between steps S4 and S5: For each data block in the Ladon domain, first calculate its standard deviation and determine the threshold according to the frequency band. Then, use the corresponding type of threshold filtering to process the data point by point to remove the focused energy of the interference wave. After processing, perform energy statistics on the data. If the interference wave energy removal rate is less than the judgment threshold, then readjust the threshold to ensure the interference suppression effect.

4. The method for suppressing linear interference in the first-arrival refraction region according to claim 2, characterized in that, Step S3 includes: 2D FFT transformation yields frequency-wavenumber domain data; Non-uniform resampling is performed according to the curvature scale parameter J; Weighted scaling with a Gaussian window function can achieve scale or orientation localization. Wrapping around the origin to locally concentrate energy; The 2D IFFT transform yields the curvelet coefficient matrix; The final output is a curve coefficient matrix with different scales and directions corresponding to low, medium and high frequency bands.

5. The method for suppressing linear interference in the first-arrival refraction region according to claim 2, characterized in that, Step S2 includes: Curved wave transform parameters: The second-generation Wrapping fast algorithm is used to design the number of scales. J , direction number L Window function bandwidth coefficient α Low-frequency bands have coarse scale and low directional resolution, while high-frequency bands have fine scale and high directional resolution, in order to adapt to the degree of fragmentation of the linear features of the interference. Radon transform parameters: A linear Radon transform is used, with the core design focusing on ray parameters. p Regularization parameters λ ray parameters p The regularization parameter λ is negatively correlated with the apparent velocity and positively correlated with the interference energy, in order to address the ill-posedness of the Radon transform and improve the focusing effect.

6. The method for suppressing linear interference in the first-arrival refraction region according to claim 5, characterized in that, Step S2 also includes: Threshold filtering parameters: Based on the energy difference between interference and effective waves, three types of thresholds are designed: hard threshold, semi-soft threshold, and soft threshold. Different threshold sizes and coefficients are used to balance the thoroughness of interference removal and the amplitude preservation of effective signals.

7. The method for suppressing linear interference in the first-arrival refraction region according to claim 2, characterized in that, Step S6 includes: First, a 2D FFT transformation is performed on the denoised curve coefficient matrix. The original position in the frequency domain is restored by inverse wrapping operation, and scale or direction inverse localization is completed by conjugate multiplication with the window function. Finally, all scale-direction data blocks are merged and transformed back to the tx domain by 2DIFFT transformation to obtain single-shot seismic data with first-arrival refraction interference removed.

8. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the steps of the linear interference suppression method for the first arrival refraction region as described in claims 1-7.

9. 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 linear interference suppression method for the first arrival refraction region as described in claims 1-7.

10. A computer device, comprising: 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 linear interference suppression method for the first arrival refraction region as described in claims 1-7.

Images