Three-dimensional gravity inversion method, system and storage medium with weakened regularization parameter

Abstract

The present invention relates to a three-dimensional gravity inversion method, system, and storage medium for weakening regularization parameters. The method comprises: generating an inversion replacement formula for weakening the regularization parameter based on a model weighting matrix and an inverse matrix of the model weighting matrix in a three-dimensional model; obtaining a first inversion formula for the three-dimensional model based on the inversion replacement formula and the original inversion formula of the three-dimensional model; and performing three-dimensional gravity inversion on the three-dimensional model based on the first inversion formula. By performing a modified replacement on the original three-dimensional gravity inversion formula, the present invention achieves three-dimensional gravity constrained inversion with weakened regularization parameters while retaining the constraints of the three-dimensional model. This improves the computational efficiency of the three-dimensional gravity inversion process while ensuring the computational accuracy of the three-dimensional gravity inversion process.

Description

Background Art

[0002] Gravity inversion is to inversely solve the three-dimensional model vector based on the observed data vector, specifically to solve high-dimensional data from low-dimensional data. Therefore, the process of inversion calculation is usually ill-posed. The purpose of gravity inversion is to construct a reasonable density model. One of the criteria for whether the density model is reasonable is whether it can fit the observed data and satisfy the constraints of prior information. Gravity inversion belongs to an ill-posed problem, and there are many regularization methods for dealing with ill-posed problems. Commonly used regularization methods include the Generalized Cross-Validation Criterion (GCV) and the L-curve criterion, etc. The above methods balance the weight relationship between the fitting error of the observed data and the norm of the three-dimensional model from different perspectives, where the regularization parameter is an important parameter for balancing the weight relationship: in terms of the fitting degree of the data, the smaller the regularization parameter, the higher the fitting degree of the observed data, but the role of model constraints cannot be exerted; from the perspective of the stability of the solution, the larger the regularization parameter, the easier it is to obtain a solution, but the fitting degree of the observed data will be lower.

[0003] It can be seen from this that in the existing inversion process, the regularization factor has an important influence on the inversion result. Usually, a suitable regularization parameter needs to be selected to possibly obtain a satisfactory result.

[0004] Therefore, if the constrained inversion of the three-dimensional model can be realized under the condition of weakening or even eliminating the regularization factor, it undoubtedly has important significance for reducing the difficulty of three-dimensional gravity inversion and improving the practicability of three-dimensional gravity inversion. Summary of the Invention

[0005] To solve the above technical problems, the present invention provides a three-dimensional gravity inversion method, system and storage medium for weakening the regularization parameter.

[0006] The technical solution of the three-dimensional gravity inversion method for weakening the regularization parameter of the present invention is as follows:

[0007] Generate an inversion replacement formula for weakening the regularization parameter according to the model weighting matrix in the three-dimensional model and the inverse matrix of the model weighting matrix;

[0008] Obtain the first inversion formula of the three-dimensional model according to the inversion replacement formula and the original inversion formula of the three-dimensional model;

[0009] Perform three-dimensional gravity inversion on the three-dimensional model according to the first inversion formula.

[0010] The beneficial effects of the three-dimensional gravity inversion method for weakening the regularization parameter of the present invention are as follows:

[0011] The method of the present invention achieves three-dimensional gravity constrained inversion with weakened regularization parameters by modifying the original three-dimensional gravity inversion formula while retaining the three-dimensional model constraints. While ensuring the calculation accuracy of the three-dimensional gravity inversion process, it also improves the calculation efficiency of the three-dimensional gravity inversion process.

[0012] Based on the above solution, the three-dimensional gravity inversion method with a weakened regularization parameter of the present invention can be further improved as follows.

[0013] Furthermore, the inversion replacement formula is: -1 W=I, where W is the model weight matrix, W -1 is the inverse matrix of the model weighting matrix, and I is the identity matrix.

[0014] Furthermore, the original inversion formula is: ρ=ρ0+(G T D -1 G+αW -1 ) -1 G T D -1 (d-Gρ0), the first inversion formula is: ρ=ρ0+(WG T D -1 G+αI) -1 WG T D -1 (d-Gρ0), where ρ is the model space of the three-dimensional model, ρ0 is the reference model, and D -1 is the diagonal data covariance matrix, and α is the regularization parameter.

[0015] Furthermore, performing three-dimensional gravity inversion on the three-dimensional model according to the first inversion formula specifically includes:

[0016] When the regularization parameter in the first inversion formula is zero, the first inversion formula evolves into the second inversion formula, which is ρ=ρ0+(WG T D -1 G) -1 WG T D -1 (d-Gρ0);

[0017] Perform three-dimensional gravity inversion on the three-dimensional model according to the second inversion formula.

[0018] Furthermore, the inversion replacement formula for weakening the regularization parameter is generated based on the model weighting matrix in the three-dimensional model and the inverse matrix of the model weighting matrix, specifically including:

[0019] When the diagonal elements of the inverse matrix of the model weighting matrix are non-zero, an inversion substitution formula for weakening the regularization parameter is generated based on the model weighting matrix and the inverse matrix of the model weighting matrix.

[0020] The technical solution of the three-dimensional gravity inversion system for weakening the regularization parameter of the present invention is as follows:

[0021] It includes: a generation module, a processing module, and an operation module;

[0022] The generation module is used for: generating an inversion substitution formula for weakening the regularization parameter according to the model weighting matrix in the three-dimensional model and the inverse matrix of the model weighting matrix;

[0023] The processing module is used for: obtaining the first inversion formula of the three-dimensional model according to the inversion substitution formula and the original inversion formula of the three-dimensional model;

[0024] The operation module is used for: performing three-dimensional gravity inversion on the three-dimensional model according to the first inversion formula.

[0025] The beneficial effects of the three-dimensional gravity inversion system for weakening the regularization parameter of the present invention are as follows:

[0026] The system of the present invention realizes three-dimensional gravity constrained inversion under the condition of weakening the regularization parameter by performing variant substitution on the original three-dimensional gravity inversion formula while retaining the three-dimensional model constraints, improving the calculation efficiency of the three-dimensional gravity inversion process while ensuring the calculation accuracy of the three-dimensional gravity inversion process.

[0027] On the basis of the above solution, the three-dimensional gravity inversion system for weakening the regularization parameter of the present invention can also be improved as follows.

[0028] Further, the inversion substitution formula is: W -1 W = I, where W is the model weighting matrix, and W -1 is the inverse matrix of the model weighting matrix, and I is the identity matrix.

[0029] Further, the original inversion formula is: ρ = ρ0 + (G T D -1 G + αW -1 ) -1 G T D -1 (d - Gρ0), and the first inversion formula is: ρ = ρ0 + (WG T D -1 G + αI) -1 WG T D -1(d - Gρ0), where ρ is the model space of the three - dimensional model, ρ0 is the reference model, and D -1 is the diagonal data covariance matrix, and α is the regularization parameter.

[0030] Furthermore, the running module is specifically configured to:

[0031] When the value of the regularization parameter in the first inversion formula is zero, the first inversion formula evolves into the second inversion formula, and the second inversion formula is ρ = ρ0+(WG T D -1 G) -1 WG T D -1 (d - Gρ0);

[0032] Perform three - dimensional gravity inversion on the three - dimensional model according to the second inversion formula.

[0033] The technical solution of a storage medium of the present invention is as follows:

[0034] The storage medium stores instructions, and when a computer reads the instructions, it causes the computer to execute the steps of the three - dimensional gravity inversion method with weakened regularization parameters of the present invention.

[0035] The technical solution of an electronic device of the present invention is as follows:

[0036] It includes a memory, a processor, and a computer program stored on the memory and executable on the processor. It is characterized in that when the processor executes the computer program, it causes the computer to execute the steps of the three - dimensional gravity inversion method with weakened regularization parameters of the present invention. Brief Description of the Drawings

[0037] Figure 1 It is a schematic flowchart of a three - dimensional gravity inversion method with weakened regularization parameters according to an embodiment of the present invention;

[0038] Figure 2 It is a three - dimensional model and a forward anomaly map in a three - dimensional gravity inversion method with weakened regularization parameters according to an embodiment of the present invention;

[0039] Figure 3 It is an inversion data fitting error map in a three - dimensional gravity inversion method with weakened regularization parameters according to an embodiment of the present invention;

[0040] Figure 4 It is a cross - sectional view of the inversion result in a three - dimensional gravity inversion method with weakened regularization parameters according to an embodiment of the present invention;

[0041] Figure 5 It is a schematic structural diagram of a three - dimensional gravity inversion system with weakened regularization parameters according to an embodiment of the present invention. Detailed implementation mode

[0042] As Figure 1 shown, a three-dimensional gravity inversion method for weakening the regularization parameter in an embodiment of the present invention includes the following steps:

[0043] S1. Generate an inversion replacement formula for weakening the regularization parameter according to the model weighting matrix in the three-dimensional model and the inverse matrix of the model weighting matrix.

[0044] Among them, the three-dimensional model is a three-dimensional model obtained by dividing the underground space.

[0045] Among them, the solution of the inversion problem of three-dimensional gravity is generally realized by minimizing the Tikhonov regularization objective function ρ(α); In the above formula, ρ(α) is the density model (three-dimensional model) of the expected three-dimensional gravity, and α is the regularization parameter; is the data fitting error objective function, is the Tikhonov regularization objective function, D -1 is the diagonal data covariance matrix (with a dimension of N×N), the elements on its diagonal are the estimated data noise variances, ρ0 is the reference model, W is the model weighting matrix (with a dimension of M×M), and when W is only the depth weighting matrix, the element W jj on its diagonal is the depth weighting function, and both M and N are positive integers.

[0046] Among them, the model weighting matrix is W, and the inverse matrix of the model weighting matrix is W -1 ; when the diagonal elements of W -1 are all non-zero, there is an inversion replacement formula: W -1 W = I, where I is the identity matrix (with a dimension of M×M). The inversion replacement formula is used to separate the regularization parameter in the original inversion formula from the inverse matrix W -1 of the model weighting matrix to weaken the influence of the regularization parameter on the three-dimensional model inversion.

[0047] It should be noted that the model weighting matrix W can be a depth weighting matrix, or a smoothness constraint matrix, a focusing constraint matrix, etc.

[0048] S2. Obtain the first inversion formula of the three-dimensional model according to the inversion replacement formula and the original inversion formula of the three-dimensional model.

[0049] Among them, the original inversion formula of the three-dimensional model is: ρ = ρ0 + (G T D -1 G + αW -1 ) -1 G T D-1 (d - Gρ0), where G is the forward kernel matrix of the three - dimensional model, G T is the transpose matrix of G, ρ is the model space of the three - dimensional model, ρ0 is the reference model, D -1 is the diagonal data covariance matrix, and α is the regularization parameter.

[0050] Specifically, by transforming the inversion replacement formula W -1 W = I to get W -1 = I / W, and substituting it into the original inversion formula: ρ = ρ0+(G T D -1 G + αW -1 ) -1 G T D -1 (d - Gρ0), the first inversion formula is obtained: ρ = ρ0+(WG T D -1 G + αI) -1 WG T D -1 (d - Gρ0).

[0051] S3. Perform three - dimensional gravity inversion on the three - dimensional model according to the first inversion formula.

[0052] Among them, by separating the regularization parameter α from W -1 , regardless of the value of the regularization parameter, the model weighting matrix W and the diagonal data covariance matrix D -1 are retained; therefore, performing three - dimensional gravity inversion according to the first inversion formula realizes the constrained inversion of three - dimensional gravity while weakening the regularization parameter.

[0053] It should be noted that the regularization parameter α is used to balance the weights between the matrix G T D -1 G and W -1 .

[0054] Preferably, the inversion replacement formula is: W -1 W = I, where W is the model weighting matrix, W -1 is the inverse matrix of the model weighting matrix, and I is the identity matrix.

[0055] Among them, both W and W -1 are diagonal matrices.

[0056] Preferably, the original inversion formula is: ρ = ρ0+(G T D -1 G + αW -1 ) -1 G T D -1(d - Gρ0), and the first inversion formula is: ρ = ρ0 + (WG T D -1 G + αI) -1 WG T D -1 (d - Gρ0), where ρ is the model space of the three - dimensional model, ρ0 is the reference model, D -1 is the diagonal data covariance matrix, and α is the regularization parameter.

[0057] Preferably, the three - dimensional gravity inversion of the three - dimensional model according to the first inversion formula specifically includes:

[0058] When the value of the regularization parameter in the first inversion formula is zero, the first inversion formula evolves into a second inversion formula, and the second inversion formula is ρ = ρ0 + (WG T D -1 G) -1 WG T D -1 (d - Gρ0);

[0059] Perform three - dimensional gravity inversion on the three - dimensional model according to the second inversion formula.

[0060] Among them, if α = 0 in the original inversion formula, the original inversion deformation formula is obtained according to the original inversion formula: ρ = ρ0 + (G T D -1 G) -1 G T D -1 (d - Gρ0). At this time, when there is no regularization parameter α, the inverse matrix W of the model weighting matrix is removed from this formula -1 , and only the diagonal data covariance matrix D -1 is retained. Therefore, if the inversion is performed according to this formula, the constrained inversion of the three - dimensional model cannot be achieved.

[0061] Specifically, when the regularization parameter α in the first inversion formula is taken as zero, the variant formula (the second inversion formula) of the first inversion formula is obtained: ρ = ρ0 + (WG T D -1 G) -1 WG T D -1 (d - Gρ0); Compared with the original inversion deformation formula, the second inversion formula removes the regularization parameter α while retaining the model weighting matrix W and the diagonal data covariance matrix D -1 ; Therefore, performing three - dimensional gravity inversion according to the second inversion formula realizes the constrained inversion of three - dimensional gravity while weakening the regularization parameter.

[0062] Preferably, generating an inversion replacement formula for weakening the regularization parameter according to the model weighting matrix and the inverse matrix of the model weighting matrix in the three-dimensional model specifically includes:

[0063] When the diagonal elements of the inverse matrix of the model weighting matrix are non-zero, generating the inversion replacement formula for weakening the regularization parameter based on the model weighting matrix and the inverse matrix of the model weighting matrix.

[0064] Among them, the condition for the establishment of the inversion replacement formula is that the diagonal elements of the inverse matrix \(W\) of the model weighting matrix are all non-zero. -1 in the are all non-zero.

[0065] To illustrate the role of the regularization parameter in the three-dimensional gravity inversion formula, a cuboid anomaly is set as the three-dimensional model. The density of this anomaly is \(0.4g / cm^3\). 3 , Figure 2 (a) is the projection position of the anomaly on the horizontal plane, and the projection is basically located in the center of the survey area. Figure 2 (c) is the projection position of the anomaly on the vertical section, and the projection position of the vertical section on the plane is located at the dotted line position in Figure 2 (a). Through calculation, the forward anomaly field of the model ( Figure 2 (b)) is obtained, and the anomaly curve of the anomaly field on the section is a single-peak anomaly ( Figure 2 (d)). The underground space is divided into \(51\times53\times31\) prismatic elements of \(100m\times100m\times100m\), and the ground observation data is located at the plane center of the prismatic body, with a total of \(51\times53\) observation data.

[0066] The model tests are respectively carried out for inversion calculations according to the original inversion formula and the first inversion formula. The number of inversion iterations is 35 times for both, and the initial models are all 0 values. The model tests only change the setting of the regularization parameter. Three kinds of parameters are selected for the regularization parameter according to the degree of data fitting. Selecting 0.1 represents the case of overfitting of data, selecting 200 represents the case of optimal fitting of data, and selecting 1000 represents the case of underfitting of data.

[0067] The data fitting error graph is used to represent the degree of difference between the inversion result and the forward field. The data fitting error graph is the difference between the forward field of the inversion model and the forward field of the theoretical model. According to the data fitting error graphs of the inversion results according to the original inversion formula ( Figure 3 (a), Figure 3 (c), Figure 3 (e)) and the data fitting error graphs of the inversion results according to the first inversion formula ( Figure 3 (b), Figure 3 (d), Figure 3(f)), it can be seen that the influence degree of the regularization parameter on the data fitting degree is the same for the two formulas, that is, the smaller the regularization parameter, the higher the fitting accuracy of the data.

[0068] To analyze the influence degree of the regularization parameter on the inversion result, the density profile anomaly maps of the inversion results are used for comparison. The projection position of the profile diagram on the plane is Figure 3 the position of the dashed line in (a). According to the inversion results calculated by the original inversion formula ( Figure 4 (a), Figure 4 (c), Figure 4 (e)), it can be seen that when the value of the regularization parameter is too small, the anomaly range of the inversion result is relatively concentrated and the anomaly value is relatively large. However, the anomaly center of the inversion result is significantly shallower than the model center (the black frame in the figure), with an obvious skin effect, indicating that the influence of the depth weighting function on the inversion result is insufficient; when the value of the regularization parameter is appropriate, the anomaly center of the inversion result coincides with the model center, but the anomaly range of the inversion result is relatively large and the anomaly value is relatively small; when the value of the regularization parameter is too large, the anomaly center of the inversion result is deeper than the model center. Although the anomaly range of the inversion result is somewhat concentrated, the anomaly value is significantly reduced.

[0069] According to the inversion results calculated by the first inversion formula ( Figure 4 (b), Figure 4 (d), Figure 4 (f)), it can be seen that when the value of the regularization parameter is too small, although the anomaly range of the inversion result is relatively large, the anomaly center of the inversion result is only slightly shallower than the model center (the black frame in the figure), without an obvious skin effect; when the value of the regularization parameter is appropriate, the anomaly center of the inversion result coincides with the model center, and the anomaly range of the inversion result is also relatively large; when the value of the regularization parameter is too large, the anomaly center of the inversion result is deeper than the model center. Although the anomaly range of the inversion result is somewhat concentrated, the anomaly value is significantly reduced, and the inversion result is basically the same as that of the traditional formula.

[0070] By comparing the inversion results of the original inversion formula and the first inversion formula, it can be found that when using the original inversion formula, the regularization parameter has a great influence on the inversion result. When the value is too small (α = 0.1), the fitting degree of the inverted data is high but the inversion result is not ideal; when the value is appropriate (α = 200), the inversion result is relatively ideal, but it will affect the fitting accuracy of the data. When using the first inversion formula, the influence of the regularization parameter on the inversion result is relatively small. When the value is too small (α = 0.1), the fitting degree of the inverted data is high, and at the same time the inversion result is also relatively ideal; when the value is appropriate (α = 200), the inversion result is relatively ideal, and it will also affect the fitting accuracy of the data. For the case where the value of the regularization parameter is too large (α = 1000), the inversion results of the two formulas are basically the same, both with low data fitting degree and poor inversion results.

[0071] In summary, in the original inversion formula, the role of the regularization parameter is to balance the weights between the data term and the model term. The selection of the regularization parameter has a great impact on the inversion result. In the first inversion formula, the regularization parameter is separated from the model constraint term and no longer adjusts the weights between the data term and the model term. Therefore, in the inversion, even when the selected regularization parameter is too small, the role of the model weighting function can still be realized. Model experiments verify that the first inversion formula improved from the original inversion formula can weaken the role of the regularization parameter and reduce the difficulty in selecting the regularization parameter during the inversion process.

[0072] The technical solution of this embodiment realizes 3D gravity constrained inversion with weakened regularization parameters by performing variant substitution on the original 3D gravity inversion formula while retaining the 3D model constraints, improving the calculation efficiency of the 3D gravity inversion process while ensuring the calculation accuracy of the 3D gravity inversion process.

[0073] As Figure 5 shown, a 3D gravity inversion system 200 for weakening the regularization parameter according to an embodiment of the present invention includes: a generation module 210, a processing module 220, and an operation module 230;

[0074] The generation module 210 is configured to: generate an inversion replacement formula for weakening the regularization parameter according to the model weighting matrix in the 3D model and the inverse matrix of the model weighting matrix;

[0075] The processing module 220 is configured to: obtain the first inversion formula of the 3D model according to the inversion replacement formula and the original inversion formula of the 3D model;

[0076] The operation module 230 is configured to: perform 3D gravity inversion on the 3D model according to the first inversion formula.

[0077] Preferably, the inversion replacement formula is: W -1 W = I, where W is the model weighting matrix, W -1 is the inverse matrix of the model weighting matrix, and I is the identity matrix.

[0078] Preferably, the original inversion formula is: ρ = ρ0 + (G T D -1 G + αW -1 ) -1 G T D -1 (d - Gρ0), and the first inversion formula is: ρ = ρ0 + (WG T D -1 G + αI) -1 WGT D -1 (d-Gρ0), where ρ is the model space of the three-dimensional model, ρ0 is the reference model, and D -1 is the diagonal data covariance matrix, and α is the regularization parameter.

[0079] Preferably, the operation module 230 is specifically used to:

[0080] When the regularization parameter in the first inversion formula is zero, the first inversion formula evolves into the second inversion formula, which is ρ=ρ0+(WG T D -1 G) -1 WG T D -1 (d-Gρ0);

[0081] Perform three-dimensional gravity inversion on the three-dimensional model according to the second inversion formula.

[0082] The technical solution of this embodiment achieves three-dimensional gravity constrained inversion with weakened regularization parameters by modifying and replacing the original three-dimensional gravity inversion formula while retaining the three-dimensional model constraints. While ensuring the calculation accuracy of the three-dimensional gravity inversion process, it also improves the calculation efficiency of the three-dimensional gravity inversion process.

[0083] For the above-mentioned parameters and steps for each module to implement corresponding functions in the three-dimensional gravity inversion system 200 with weakened regularization parameters in this embodiment, reference can be made to the parameters and steps in the embodiment of the three-dimensional gravity inversion method with weakened regularization parameters above, and no further details will be given here.

[0084] A storage medium provided by an embodiment of the present invention includes: instructions stored in the storage medium, and when a computer reads the instructions, the computer executes the steps of a three-dimensional gravity inversion method with a weakened regularization parameter. For details, please refer to the parameters and steps in the embodiment of the three-dimensional gravity inversion method with a weakened regularization parameter above, which will not be repeated here.

[0085] Computer storage media such as USB flash drives, mobile hard drives, etc.

[0086] An embodiment of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. The electronic device is characterized in that when the processor executes the computer program, the computer is caused to execute steps of a three-dimensional gravity inversion method with a weakened regularization parameter. For details, reference may be made to the various parameters and steps in the embodiment of the three-dimensional gravity inversion method with a weakened regularization parameter described above, which will not be elaborated upon here.

[0087] Those skilled in the art of the present invention know that the present invention can be implemented as a method, a system, a storage medium, and an electronic device.

[0088] Therefore, the present invention can be specifically implemented in the following forms, namely: it can be completely hardware, or completely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software, which is generally referred to as "circuit", "module" or "system" in this article. In addition, in some embodiments, the present invention can also be implemented in the form of a computer program product in one or more computer-readable media, and the computer-readable media contains computer-readable program code. Any combination of one or more computer-readable media can be used. The computer-readable media can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (non-exhaustive list) of the computer-readable storage medium include: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this document, the computer-readable storage medium can be any tangible medium that contains or stores a program, and the program can be used by or in combination with an instruction execution system, apparatus, or device. Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A three-dimensional gravity inversion method for weakening regularization parameters, characterized in that, Comprising: Generating an inversion replacement formula for weakening the regularization parameter based on the model weighting matrix in the three-dimensional model and the inverse matrix of the model weighting matrix; Obtaining a first inversion formula of the three-dimensional model according to the inversion replacement formula and the original inversion formula of the three-dimensional model; Performing three-dimensional gravity inversion on the three-dimensional model according to the first inversion formula; The inversion replacement formula is: W -1 W = I, where W is the model weighting matrix, and W -1 is the inverse matrix of the model weighting matrix, and I is the identity matrix; The original inversion formula is: ρ = ρ0 + (G T D -1 G + αW -1 ) -1 G T D -1 (d - Gρ0), and the first inversion formula is: ρ = ρ0 + (WG T D -1 G + αI) -1 WG T D -1 (d - Gρ0), where ρ is the model space of the three-dimensional model, ρ0 is the reference model, D -1 is the diagonal data covariance matrix, α is the regularization parameter, G is the forward kernel matrix of the three-dimensional model, and G T is the transpose matrix of G.

2. The three-dimensional gravity inversion method for weakening the regularization parameter according to claim 1, wherein The performing three-dimensional gravity inversion on the three-dimensional model according to the first inversion formula specifically includes: When the regularization parameter in the first inversion formula takes a value of zero, the first inversion formula evolves into a second inversion formula, and the second inversion formula is ρ = ρ0 + (WG T D -1 G) -1 WG T D -1 (d - Gρ0); Performing three-dimensional gravity inversion on the three-dimensional model according to the second inversion formula.

3. The three-dimensional gravity inversion method for weakening the regularization parameter according to claim 1 or 2, characterized in that The generating an inversion replacement formula for weakening the regularization parameter based on the model weighting matrix in the three-dimensional model and the inverse matrix of the model weighting matrix specifically includes: When the diagonal elements of the inverse matrix of the model weighting matrix are non-zero, generating the inversion replacement formula for weakening the regularization parameter based on the model weighting matrix and the inverse matrix of the model weighting matrix.

4. A three-dimensional gravity inversion system for weakening the regularization parameter, characterized in that, Comprising: A generating module, a processing module and a running module; The generating module is configured to: generate an inversion replacement formula for weakening the regularization parameter based on the model weighting matrix in the three-dimensional model and the inverse matrix of the model weighting matrix; The processing module is configured to: obtain a first inversion formula of the three-dimensional model according to the inversion replacement formula and the original inversion formula of the three-dimensional model; The running module is configured to: perform three-dimensional gravity inversion on the three-dimensional model according to the first inversion formula; The inversion replacement formula is: W -1 W = I, where W is the model weighting matrix, and W -1 is the inverse matrix of the model weighting matrix, and I is the identity matrix; The original inversion formula is: ρ = ρ0 + (G T D -1 G + αW -1 ) -1 G T D -1 (d - Gρ0), and the first inversion formula is: ρ = ρ0 + (WG T D -1 G + αI) -1 WG T D -1 (d - Gρ0), where ρ is the model space of the three-dimensional model, ρ0 is the reference model, D -1 is the diagonal data covariance matrix, α is the regularization parameter, G is the forward kernel matrix of the three-dimensional model, and G T is the transpose matrix of G.

5. The three-dimensional gravity inversion system for weakening the regularization parameter according to claim 4, wherein The running module is specifically configured to: When the regularization parameter in the first inversion formula takes a value of zero, the first inversion formula evolves into a second inversion formula, and the second inversion formula is ρ = ρ0 + (WG T D -1 G) -1 WG T D -1 (d - Gρ0); Perform three-dimensional gravity inversion on the three-dimensional model according to the second inversion formula.

6. A storage medium, characterized in that, Instructions are stored in the storage medium, and when a computer reads the instructions, the computer is caused to execute the three-dimensional gravity inversion method for weakening the regularization parameter according to any one of claims 1 to 3.

Images