Short-wavelength residual static correction method based on target reconstruction algorithm

Abstract

The invention provides a short-wavelength residual static correction method based on a target reconstruction algorithm, and belongs to the technical field of seismic data processing. Comprising the following steps: modeling seismic data, expressing the seismic data as a form of a plurality of waves, conforming to a target sparse model, and assuming that the target sparse model contains a plurality of targets and is accompanied by noise; target decomposition is carried out on the seismic data containing the noise and reconstruction is carried out by adopting an iteration method, one target is reconstructed each time, and residual errors after decomposition are continuously decomposed until reconstruction of all targets is completed; the minimization problem of the target reconstruction process is converted into a regularization optimization problem with constraint conditions; and solving a quadratic optimization problem about the target to obtain a minimum value depending on a shift vector, and completing short-wavelength residual static correction by solving the maximum weight from the vertex 1 to the vertex N + 1 in the K approximate graph and reconstructing the shift vector. The invention provides the residual static correction method with robustness in combination with the properties of the seismic data under the condition that the calculation cost is not increased.

Description

Technical Field

[0001] This invention relates to the field of seismic data processing technology, and in particular to a short-wavelength residual static correction method based on a target reconstruction algorithm. Background Technology

[0002] With the continuous advancement of seismic exploration technology, the processing of seismic data in complex exploration areas, especially static correction in areas with complex surface conditions, has attracted widespread attention. Seismic data is affected by surface undulations and near-surface weathering layers, leading to time deviations in wave propagation. Static correction for long wavelengths is generally related to velocity structure and can usually be solved by refracted wave static correction or tomographic static correction. However, for short wavelengths and some mid-wavelengths, residual static correction is necessary. Residual static correction can eliminate measurement and observation errors as well as correct errors in near-surface velocity models and the static correction methods themselves, significantly improving the imaging quality of stacked profiles.

[0003] Currently, traditional short-wavelength static correction methods rely on the surface consistency assumption, which assumes that the near-surface ray path is vertical. In this case, the residual time difference depends only on the positions of the shot point and receiver on the ground, without considering their offset. This assumption is valid in most practical cases because the surface consistency assumption is equivalent to a simple near-surface model, and this assumption plays a crucial role in ensuring the stability of the residual static correction and avoiding distortion and misalignment of the phase axis. Most surface consistency methods use the common midpoint (CMP) gather of the dynamic correction (NMO) to estimate the residual time difference. Surface consistency methods assume that the phase axis is aligned after dynamic correction; if not, then the residual time difference or residual NMO is the main cause. However, in cases of complex reflections, dynamic correction cannot flatten all phase axes; and in low signal-to-noise ratio (SNR) conditions, it is difficult to accurately determine the phase axis with weak amplitude. Since migration considers complex situations such as tilted reflection interfaces and discontinuity diffraction, it can accurately reposition reflected waves, and the superposition also enhances the SNR of the migrated gather. Therefore, the residual time difference can also be calculated based on the migrated gather.

[0004] Furthermore, when near-surface weathering affects the coherence of seismic data, rank-based methods can be used for residual static correction. For seismic data with residual time differences, coherence decreases, leading to a slow decay of singular values. If data redundancy is utilized and the influence of weathering is corrected, a low-rank matrix needs to be estimated because the desired data without residual time differences has a low-rank property. Seismic data can be approximated by a low-rank matrix through the transform domain, for example, when it is in matrix form in the center-point offset domain; another method for calculating rank is to use sparsity to compute residual time differences. Considering reflection information in seismic exploration, recorded seismic data can be considered to have a sparse structure. Furthermore, if reflections are perfectly aligned, the data is generally low-rank.

[0005] Therefore, providing a robust residual static correction method that incorporates the properties of seismic data itself without increasing computational costs is of significant practical importance. Summary of the Invention

[0006] To overcome the shortcomings of the prior art, the purpose of this invention is to provide a short-wavelength residual static correction method based on a target reconstruction algorithm. This method provides a robust residual static correction method that combines the properties of seismic data itself without increasing computational costs. It can effectively predict residual static correction values, improve the quality of stacked profile imaging, and has high computational efficiency and a wide range of applicable data.

[0007] To achieve the above objectives, the present invention provides the following solution:

[0008] A short-wavelength residual static correction method based on a target reconstruction algorithm includes the following steps:

[0009] S1. Model the seismic data, based on the fact that the seismic data is represented as multiple waves, conforming to the target sparse model, and assuming that it contains multiple targets and is accompanied by noise;

[0010] S2. Based on the constructed target sparse model, the noisy seismic data is decomposed into targets and reconstructed using an iterative method. After reconstructing one target at a time, the residuals after decomposition are further decomposed until all targets are reconstructed.

[0011] S3. Based on the fact that the total change of the objective follows physical laws and is bounded, the minimization problem of the objective reconstruction process in step S2 is transformed into a regularized optimization problem with constraints.

[0012] S4. Solve the quadratic optimization problem about the target. Analytical solution obtains the minimum value dependent on the shift vector. By solving the maximum weight from vertex 1 to vertex N+1 in the K approximation graph and reconstructing the shift vector, the short-wavelength residual static correction is completed.

[0013] Preferably, in S1, the target sparse model includes:

[0014] Seismic data are represented as a real matrix A∈R M×N Where M is the number of time sampling points and N is the number of seismic traces;

[0015] The established target sparse model is as follows:

[0016]

[0017] In the formula, L is the target quantity, which satisfies L << min(M, N); For the l-th objective; λ l Let U be the shift vector of the l-th target.l Let be the target matrix of the l-th target, and NOISE be the noise term.

[0018] Preferably, in S2, target decomposition is performed on the noisy seismic data and iterative reconstruction is carried out, including:

[0019] Based on the constructed target sparse model, noisy seismic data is input and target decomposition is performed. After decomposition, the data is reconstructed using an iterative method.

[0020] Reconstruct a target at a time and solve a minimization problem:

[0021]

[0022] In the formula, A is the current decomposition data matrix in S2, U is the target matrix to be solved, λ is the target shift vector to be solved, and ||·|| F The Frobenius norm of the matrix;

[0023] Update residuals after reconstruction:

[0024] R = AS λ (U);

[0025] Continue to decompose the residual R until all objectives are reconstructed.

[0026] Preferably, in S3, the total change of the target follows physical laws and is bounded, that is:

[0027] Bounded;

[0028] In the formula, U :k is the k-th column of the target matrix.

[0029] Preferably, in S3, the minimization problem is transformed into an optimization problem with regularization terms, that is:

[0030]

[0031] The shift vector must satisfy:

[0032] |λ k -λ k+1 |≤C,k=1,…,N-1;

[0033] In the formula, S -λ is the shift operator, μ is the regularization parameter, and C is the Lipschitz constant.

[0034] Preferably, in S3, the process of solving the quadratic optimization problem is as follows:

[0035] Assume that the shift vector λ is a given value, and then analytically solve the quadratic minimization problem for the target U to obtain the minimum value that depends on the shift vector λ:

[0036]

[0037] Solve the shift vector λ by minimizing the minimum value:

[0038]

[0039] Preferably, in S4, the Lipschitz constraint that the shift vector reconstruction needs to satisfy is:

[0040] |λ k -λ k+1 |≤C, k = 1, …, N - 1;

[0041] In the formula, λ k is the k-th component of the shift vector.

[0042] Preferably, in S4, the shift vector is reconstructed by solving the maximum weight sum from vertex 1 to vertex N + 1 in the K approximate graph, that is:

[0043] Construct a K approximate graph, only retain the number that satisfies the adjacent constraint and the Lipschitz constraint condition, considering that:

[0044] -C|j - k|≤λ j -λ k ≤C|j - k|;

[0045] Therefore, there are 2C|j - k| + 1 different values for each summand in the process of solving the minimum value;

[0046] By restricting |j - k|≤N and K << N, the number of possible values will be reduced, thereby reducing the computational complexity and ensuring a small optimization error;

[0047] Finally, output the reconstruction result, that is, the shift vector λ corresponding to the optimal path is the remaining correction value, and the short - wavelength residual static correction is completed;

[0048] Among them, λ j is the shift vector component with the serial number j, and λ k is the shift vector component with the serial number k.

[0049] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0050] (1) By constructing a target sparse model, the present invention iteratively decomposes the target components in the seismic data, accurately separates noise from effective signals, significantly improves the prediction accuracy of the residual static correction value compared with traditional methods, effectively flattens the in-phase axis in complex reflection scenarios, and greatly improves the quality of the superimposed profile.

[0051] (2) In the target signal separation stage, this invention uses regularization optimization as a constraint and leverages the difference in energy distribution between noise and the effective target signal in the sparse domain to drive iterative shrinkage of the residual term, gradually stripping noise energy from the seismic data. Simultaneously, relying on the regularization mechanism to suppress noise interference, a noise-resistant environment is created for target signal reconstruction. Thus, even if the seismic data contains a certain amount of noise, the target signal can still be accurately extracted for residual static correction calculations, significantly enhancing the robustness and stability of the static correction process compared to traditional methods.

[0052] (3) This invention reduces the data size of a single reconstruction by iterative decomposition and uses graph approximation to solve the shift vector, i.e. the residual static correction value, thereby improving computational efficiency while ensuring reconstruction accuracy. At the same time, because it does not rely on the surface consistency assumption, it breaks the limitation of complex geological conditions and can be adapted to data with different noise levels and diverse geological scenarios, thus having a wider range of applications. Attached Figure Description

[0053] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0054] Figure 1 This is a flowchart of a short-wavelength residual static correction method based on a target reconstruction algorithm according to the present invention;

[0055] Figure 2 The velocity model for offsetting provided by this invention;

[0056] Figure 3 This is a seismic data map provided in Embodiment 1 of the present invention; wherein, Figure 3 (a) in the image is an uncorrected seismic data map. Figure 3 (b) in the figure is a seismic data map corrected by the method of the present invention;

[0057] Figure 4 This is a comparison chart of post-stack data provided in Embodiment 1 of the present invention; wherein, Figure 4 (a) in the figure is the uncorrected post-stack data plot. Figure 4 (b) in the figure is a diagram of the post-stack data corrected by the method of the present invention;

[0058] Figure 5 This is a comparison chart of the predicted residual static correction value and the actual residual static correction value provided in Embodiment 1 of the present invention;

[0059] Figure 6 This is a seismic data map containing Gaussian noise with a mean of 0 and a variance of 0.05, provided in Embodiment 2 of the present invention; wherein, Figure 6 (a) in the image is an uncorrected seismic data map. Figure 6 (b) in the figure is a seismic data map corrected by the method of the present invention;

[0060] Figure 7 This is a post-stacked data image containing Gaussian noise with a mean of 0 and a variance of 0.05, provided in Embodiment 2 of the present invention; wherein, Figure 7 (a) in the figure is the uncorrected post-stack data plot. Figure 7 (b) in the figure is a diagram of the post-stack data corrected by the method of the present invention;

[0061] Figure 8 This is a comparison chart of the predicted residual static correction value and the actual residual static correction value for seismic data containing Gaussian noise with a mean of 0 and a variance of 0.05, provided in Embodiment 2 of the present invention.

[0062] Figure 9 This is a seismic data map containing Gaussian noise with a mean of 0 and a variance of 0.1, provided in Embodiment 3 of the present invention; wherein, Figure 9 (a) in the image is an uncorrected seismic data map. Figure 9 (b) in the figure is a seismic data map corrected by the method of the present invention;

[0063] Figure 10 This is a post-stacked data image containing Gaussian noise with a mean of 0 and a variance of 0.1, provided in Embodiment 3 of the present invention; wherein, Figure 10 (a) in the figure is the uncorrected post-stack data plot. Figure 10 (b) in the figure is a diagram of the post-stack data corrected by the method of the present invention;

[0064] Figure 11 This is a comparison chart of the predicted residual static correction value and the actual residual static correction value for seismic data containing Gaussian noise with a mean of 0 and a variance of 0.1, provided in Embodiment 3 of the present invention. Detailed Implementation

[0065] 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 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 are within the scope of protection of the present invention.

[0066] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0067] like Figure 1 As shown, this invention provides a short-wavelength residual static correction method based on a target reconstruction algorithm, comprising the following steps:

[0068] S1. Model the seismic data, based on the fact that the seismic data is represented as multiple waves, conforming to the target sparse model, and assuming that it contains multiple targets and is accompanied by noise;

[0069] S2. Based on the constructed target sparse model, the noisy seismic data is decomposed into targets and reconstructed using an iterative method. After reconstructing one target at a time, the residuals after decomposition are further decomposed until all targets are reconstructed.

[0070] S3. Based on the fact that the total change of the objective follows physical laws and is bounded, the minimization problem of the objective reconstruction process in step S2 is transformed into a regularized optimization problem with constraints.

[0071] S4. Solve the quadratic optimization problem about the target. Analytical solution obtains the minimum value dependent on the shift vector. By solving the maximum weight from vertex 1 to vertex N+1 in the K approximation graph and reconstructing the shift vector, the short-wavelength residual static correction is completed.

[0072] In S1, the target sparse model includes:

[0073] For seismic data A∈R M×N A target-sparse model is constructed, assuming that the data is target-sparse, containing only a small number of different targets, that the targets follow basic physical laws, that their changes between two observations are structured, that the movement position information satisfies Lipschitz continuity, and that the Lipschitz constant is related to the target's maximum velocity, thus the total variation is bounded. Furthermore, since seismic data can be represented as multiple waves, it is further assumed that the data contains multiple targets and is accompanied by noise. Here, M represents the number of time sampling points, and N represents the number of seismic traces.

[0074] The established target sparse model is as follows:

[0075]

[0076] In the formula, L is the number of targets. Since it is assumed that A is sparse, the number of targets will be very small compared with the data size, that is, L << min(M, N). For the l-th objective; λ l Let U be the shift vector of the l-th target. l Let be the target matrix of the l-th target, and NOISE be the noise term.

[0077] In S2, target decomposition is performed on noisy seismic data, and iterative methods are used for reconstruction, including:

[0078] Based on the constructed target sparse model, the seismic data A, which may contain noise, is decomposed into multiple targets S. λ (U) uses an iterative method to reconstruct only one target at a time; after a target is reconstructed from the given data, the residual is decomposed until all targets are reconstructed.

[0079] During the iteration process, for each reconstructed objective, it is necessary to determine the most suitable objective for the current data, i.e., to solve a minimization problem:

[0080]

[0081] In the formula, A is the current decomposition data matrix in S2, U is the target matrix to be solved, λ is the target shift vector to be solved, and ||·|| F The Frobenius norm of the matrix;

[0082] Update residuals after reconstruction:

[0083] R = AS λ (U);

[0084] Continue to decompose the residual R until all objectives are reconstructed.

[0085] Furthermore, the minimization problem contains two constraints: the total change in the objective follows physical laws and is bounded, namely:

[0086] Bounded;

[0087] In the formula, U :k Let λ be the k-th column of the target matrix. Furthermore, λ is Lipschitz continuous, but in practice, it is difficult to obtain prior information about the total change, so a penalty term is used to include constraints. In addition, considering the properties of the shift operator,

[0088] The minimization problem is transformed into an optimization problem with regularization terms, i.e.:

[0089]

[0090] The shift vector must satisfy:

[0091] |λ k -λ k+1 |≤C,k=1,…,N-1;

[0092] In the formula, S -λ is the shift operator, μ is the regularization parameter, and C is the Lipschitz constant.

[0093] In S3, the process of secondary optimization solution is as follows:

[0094] Assume that the shift vector λ is a given value, and then analytically solve the quadratic minimization problem for the target U to obtain the minimum value dependent on the shift vector λ:

[0095]

[0096] Solve the shift vector λ by minimizing the minimum value:

[0097]

[0098] In S4, the Lipschitz constraint that the shift vector reconstruction needs to satisfy is:

[0099] |λ k -λ k+1 |≤C, k = 1, …, N - 1;

[0100] In the formula, λ k is the k-th component of the shift vector.

[0101] In S4, reconstruct the shift vector by solving the maximum weight from vertex 1 to vertex N + 1 in the K approximate graph, that is:

[0102] Construct the K approximate graph, only retain the number that satisfies the adjacent constraint and the Lipschitz constraint conditions, considering that:

[0103] -C|j - k|≤λ j -λ k ≤C|j - k|;

[0104] Therefore, each addend in the process of solving the minimum value has 2C|j - k| + 1 different values;

[0105] By restricting |j - k|≤N and K << N, the number of possible values will be reduced, thereby reducing the computational complexity and ensuring a small optimization error;

[0106] Finally, output the reconstruction result, that is, the shift vector λ corresponding to the optimal path is the remaining static correction value, and the short-wavelength remaining static correction is completed;

[0107] where, λ j is the shift vector component with the serial number j, and λ k is the shift vector component with the serial number k.

[0108] According to the above content, the present invention will be further described through the following specific embodiments.

[0109] Embodiment 1

[0110] In this embodiment, an algorithm based on target reconstruction (ORKA) is used for experimental verification. The experiment relies on the Kirchhoff depth migration velocity model (refer to...). Figure 2 The generated data, the velocity model contains 181 shot points with a shot point spacing of 12.5 meters. Offset domain common imaging point gathers (ODCIGs) were obtained through Kirchhoff depth migration, each OCCIG containing 20 channels, with a maximum offset distance of 2100 meters. To simulate the short-wavelength residual static correction problem, a 25-meter vertical perturbation was manually and randomly added to each channel of each OCCIG.

[0111] Figure 3 (a) and (b) in the figure present the common imaging point gather (CIG) with residual static correction values, and the results after correction by the ORKA algorithm, which can be used to visually compare the changes in the in-phase axis before and after correction; Figure 4 (a) and (b) in the figure show the effect of all CIGs before and after correction and are used to evaluate the overall quality of the stacked profile; Figure 5 This compares the residual static correction values ​​predicted by the algorithm with those randomly added. The experimental results show that after correction using the ORKA algorithm, the in-phase axes of the gather are more continuous, and the quality of the superimposed profile is significantly improved. This verifies that the algorithm can accurately predict residual static correction values ​​and has good short-wavelength residual static correction performance.

[0112] Example 2

[0113] In this embodiment, the Target Reconstruction Algorithm (ORKA) was used for the experiment. A velocity model for Kirchhoff depth migration (181 shot points, shot-to-shot spacing 12.5 meters) was used. This model was used to generate offset domain common imaging point gathers (ODCIGs) via Kirchhoff depth migration. Each OCCIG contained 20 channels, with a maximum offset distance of 2100 meters. To simulate the short-wavelength problem, a random vertical perturbation of 25 meters was applied to each channel of each OCCIG. Simultaneously, to test the robustness of the algorithm, Gaussian noise with a mean of 0 and a variance of 0.05 was added to the gathers.

[0114] Figure 6 (a) and (b) in the figure present the CIG with residual static correction values ​​and the results after ORKA algorithm correction. Figure 7 In the image, (a) and (b) show the combined effect of all CIG corrections before and after. Figure 8 The image shows a comparison between the residual static correction values ​​predicted by the algorithm and the randomly added residual static correction values. It can be seen that after correction by the ORKA algorithm, the phase axes of the gather are more continuous, the quality of the stacked profile is improved, the algorithm can accurately predict the residual static correction values, demonstrating robustness and the ability to handle noisy seismic data.

[0115] Example 3

[0116] This embodiment also employs the Target Reconstruction Algorithm (ORKA). Based on the same Kirchhoff depth migration velocity model and generated ODCIGs (181 shot points, 12.5-meter spacing, 20 channels per ODCIG, and a maximum offset distance of 2100 meters) as in Embodiment 2, a 25-meter perturbation is randomly applied to each channel of each ODCIG to simulate a short-wavelength problem, and a Gaussian noise test algorithm with a mean of 0 and a variance of 0.1 is added. Figure 9 Images (a) and (b) show the CIG with residual static correction values ​​and the results after algorithmic correction, respectively. Figure 10 In the image, (a) and (b) show the combined effects before and after all CIG corrections, respectively. Figure 11 This paper compares the residual static correction values ​​predicted by the algorithm with those of randomly added values. The results show that after correction by the ORKA algorithm, the continuity of the phase axes of the gather is enhanced, the quality of the stacked profile is significantly improved, and the algorithm can predict the residual static correction values ​​with basic accuracy under higher noise interference, further verifying the robustness of the algorithm and its applicability to noisy seismic data scenarios.

[0117] Therefore, the short-wavelength residual static correction method based on the target reconstruction algorithm described above provides a robust residual static correction method that combines the properties of seismic data itself without increasing computational costs. It can effectively predict residual static correction values, improve the quality of stacked profile imaging, and has high computational efficiency and a wide range of applicable data.

[0118] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0119] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A short-wavelength residual static correction method based on a target reconstruction algorithm, characterized in that, It includes the following steps: S1. Model the seismic data, which is expressed as multiple waves in accordance with the target sparse model, and assume that it contains multiple targets accompanied by noise. S2. Based on the constructed target sparse model, decompose the noisy seismic data into targets and use an iterative method for reconstruction. After reconstructing one target each time, continue to decompose the residual after decomposition until all targets are reconstructed. S3. Based on the fact that the total variation of the target follows physical laws and is bounded, transform the minimization problem in the target reconstruction process in step S2 into a regularized optimization problem with constraint conditions. S4. Solve the quadratic optimization problem for the target, analytically obtain the minimum value depending on the shift vector, and complete the short-wavelength residual static correction by solving the maximum weight sum from vertex 1 to vertex N + 1 in the K approximate graph and reconstructing the shift vector.

2. The short-wavelength residual static correction method based on the target reconstruction algorithm according to claim 1, characterized in that, In S1, the target sparse model includes: Seismic data are represented as a real matrix A∈R M×N Where M is the number of time sampling points and N is the number of seismic traces; The established target sparse model is: In the formula, L is the target quantity, which satisfies L << min(M, N); For the l-th objective; λ l Let U be the shift vector of the l-th target. l Let be the target matrix of the l-th target, and NOISE be the noise term.

3. The short-wavelength residual static correction method based on the target reconstruction algorithm according to claim 1, characterized in that, In S2, decomposing the noisy seismic data into targets and using an iterative method for reconstruction includes: Based on the constructed target sparse model, input the noisy seismic data and perform target decomposition, and then perform reconstruction through an iterative method after decomposition. Each time one target is reconstructed, solve the minimization problem: In the formula, A is the current decomposition data matrix in S2, U is the target matrix to be solved, λ is the target shift vector to be solved, and ||·|| F The Frobenius norm of the matrix; Update the residual after reconstruction: R=A-S λ (U); Continue to decompose the residual R until all targets are reconstructed.

4. The short-wavelength residual static correction method based on the target reconstruction algorithm according to claim 1, characterized in that, In S3, the total variation of the target follows physical laws and is bounded, that is: In the formula, U :k is the k-th column of the target matrix.

5. The short-wavelength residual static correction method based on the target reconstruction algorithm according to claim 4, characterized in that, In S3, transform the minimization problem into an optimization problem with a regularization term, that is: The shift vector needs to satisfy: |l k -l k+1 |≤C,k=1,…,N-1; In the formula, S -λ is the shift operator, μ is the regularization parameter, and C is the Lipschitz constant.

6. The short-wavelength residual static correction method based on the target reconstruction algorithm according to claim 5, characterized in that, In S3, the process of quadratic optimization solution is: Assume that the shift vector λ is a given value, and then analytically solve the quadratic minimization problem for the target U to obtain the minimum value depending on the shift vector λ: Solve the shift vector λ by minimizing the minimum value:

7. The short-wavelength residual static correction method based on the target reconstruction algorithm according to claim 1, characterized in that, In S4, the Lipschitz constraint that the shift vector reconstruction needs to satisfy is: |l k -l k+1 |≤C,k=1,…,N-1; In the formula, λ k This is the k-th component of the shift vector.

8. The short-wavelength residual static correction method based on the target reconstruction algorithm according to claim 1, characterized in that, In S4, reconstruct the shift vector by solving the maximum weight sum from vertex 1 to vertex N + 1 in the K approximate graph, that is: Construct the K approximate graph, only retain the number that satisfies the adjacent constraint and Lipschitz constraint conditions, considering that: -C|jk|≤λ j −λ k ≤C|jk|| Therefore, there are 2C|j - k| + 1 different values for each addend in the process of solving the minimum value; By restricting |j - k| ≤ N and K << N, the number of possible values can be reduced, thereby reducing the computational complexity and ensuring a small optimization error; Finally, output the reconstruction result, that is, the shift vector λ corresponding to the optimal path is the residual static correction value, and complete the short-wavelength residual static correction. Where, λ j Let λ be the shift vector component with index j. k Let k be the shift vector component of the sequence number k.

Images