Efficient joint inversion method and device for large-area aviation gravity and magnetic multi-parameter data
By adopting random algorithms and alternating direction strategies in the joint inversion of large-area aerial heavy-magnetic multi-parameter data, the problem of low computing efficiency is solved, efficient data inversion is achieved, and its application in large-area exploration is promoted.
Patent Information
- Application Number
- CN202411189039.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-28
- Publication Date
- 2024-12-06
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing joint inversion method of large-area aerial heavy magnetism multi-parameter data has low computational efficiency, which hinders its application in exploration of large-area regional scenarios.
The first random algorithm is used to invert each parameter data, obtain the model perturbation terms, and minimize the cross-gradient terms through the target algorithm, obtain the model correction terms until the iteration stop condition is met, and random singular values are used to decompose and alternate directions. Strategy improves computing efficiency.
It realizes efficient calculation of aviation heavy magnetic multi-parameter data in large-area exploration, improves the efficiency of joint inversion, and promotes its application in actual exploration.
Smart Images

Figure CN119087531A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of geophysics, and in particular to a high-efficiency joint inversion method and device for large-area aerogravimetric and magnetic multi-parameter data. Background Art
[0002] Airborne gravity and magnetic exploration has the advantages of high acquisition efficiency and low detection cost, and can effectively overcome environmental noise interference and measure high-precision and high-resolution airborne gravity and magnetic multi-parameter data. Joint inversion using airborne gravity and magnetic multi-parameter data is one of the important technical means for multi-parameter and high-precision detection. However, the current large-area airborne gravity and magnetic multi-parameter data joint inversion method still generally has the problem of low computational efficiency. Due to the low computational efficiency of the current large-area airborne gravity and magnetic multi-parameter data joint inversion method, it seriously hinders the application of airborne gravity and magnetic multi-parameter data joint inversion method in large-area regional exploration scenarios. Summary of the invention
[0003] The object of the present invention is to provide a method and device for efficient joint inversion of large-area aerogravimetric and magnetic multi-parameter data, which can improve the efficiency of joint inversion of large-area aerogravimetric and magnetic multi-parameter data.
[0004] To achieve the above object, the present invention provides an efficient joint inversion method for large-area aerogravimetric and magnetic multi-parameter data, comprising:
[0005] Using a first random algorithm, each parameter data in the large-area aerogravimetric and magnetic multi-parameter data is inverted respectively to obtain a model disturbance term of each parameter data;
[0006] Adopting the target algorithm, based on the model disturbance terms of different parameter data, the cross gradient term is minimized to obtain the model correction term of each parameter data;
[0007] Based on the model correction term of each parameter data, the model of each parameter data is updated until the iterative stopping condition is met.
[0008] In one embodiment of the present invention, the first random algorithm is used to invert each parameter data in the large-area aerogravimetric and magnetic multi-parameter data to obtain the model disturbance term of each parameter data, including:
[0009] For each of the parameter data, transforming the inversion objective function of the parameter data from a function of a model vector of the parameter data to a first function of a first vector;
[0010] Decomposing the first matrix based on a random singular value decomposition algorithm to obtain the first vector;
[0011] Based on the first vector and the model weighting matrix of the parameter data, a model disturbance term of the parameter data is obtained.
[0012] In one embodiment of the present invention, the target algorithm is used to minimize the cross gradient term based on the model disturbance term of different parameter data to obtain the model correction term of each parameter data, including:
[0013] For different parameter data, converting the formula for minimizing the cross gradient term into a minimum two-norm form;
[0014] Based on the formula for minimizing the cross gradient term in the form of a least squares norm, the model correction term for the different parameter data is obtained.
[0015] In one embodiment of the present invention, the formula for minimizing the cross gradient term based on the least two norm form to obtain the model correction term for the different parameter data includes:
[0016] Based on the formula for minimizing the cross-gradient term in the form of the least square norm, an alternating direction strategy is adopted to process the cross-gradient term;
[0017] Based on the cross-gradient term processed by the alternating direction strategy, the formula for minimizing the cross-gradient term in the form of the least square norm is solved to obtain the model correction term of the different parameter data.
[0018] In one embodiment of the present invention, the formula for minimizing the cross gradient term based on the least two norm form uses an alternating direction strategy to process the cross gradient term, including:
[0019] A direction is selected as a target direction in sequence according to a preset order, and in minimizing the cross gradient term in each iteration, the cross gradient term is processed based on the target direction.
[0020] In one embodiment of the present invention, the cross gradient term processed based on the alternating direction strategy, solving the formula for minimizing the cross gradient term in the form of the least square norm to obtain the model correction term for the different parameter data, includes:
[0021] Decomposing the second matrix based on a random singular value decomposition algorithm to obtain a second vector;
[0022] Based on the second vector and the model disturbance term of the different parameter data, a model correction term of the different parameter data is obtained.
[0023] In one embodiment of the present invention, a highly efficient joint inversion device for large-area aerogravimetric and magnetic multi-parameter data comprises:
[0024] An independent inversion module is used to adopt a first random algorithm to invert each parameter data in the large-area aerogravimetric and magnetic multi-parameter data, and obtain a model disturbance term of each parameter data;
[0025] A cross gradient term minimization module is used to adopt a target algorithm to minimize the cross gradient term based on the model disturbance term of different parameter data to obtain the model correction term of each parameter data;
[0026] An updating module is used to update the model of each parameter data based on the model correction item of each parameter data until the iterative stopping condition is met.
[0027] In one embodiment of the present invention, an electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps of the efficient joint inversion method for large-area aerogravimetric and multi-parameter data as described in any one of the above are implemented.
[0028] In one embodiment of the present invention, a non-transitory computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of any of the above-mentioned methods for efficient joint inversion of large-area aerogravimetric and multi-parameter data.
[0029] In one embodiment of the present invention, a computer program product includes a computer program, which, when executed by a processor, implements the steps of any of the above-mentioned methods for efficient joint inversion of large-area airborne gravity and magnetic multi-parameter data.
[0030] Compared with the prior art, the efficient joint inversion method and device for large-area airborne gravity and magnetic multi-parameter data according to the present invention has the beneficial effect that, by adopting efficient strategies in both independent inversion and cross-gradient term minimization, efficient calculation of joint inversion of airborne gravity and magnetic multi-parameter data in a sequential manner can be achieved, the efficiency of joint inversion of large-area airborne gravity and magnetic multi-parameter data is improved, and the application of joint inversion of airborne gravity and magnetic multi-parameter data in large-area actual exploration can be realized. BRIEF DESCRIPTION OF THE DRAWINGS
[0031] Figure 1 This is one of the flow diagrams of an efficient joint inversion method for large-area aerogravimetric and magnetic multi-parameter data according to an embodiment of the present invention;
[0032] Figure 2 This is the second flow chart of the efficient joint inversion method of large-area aerogravimetric and magnetic multi-parameter data according to one embodiment of the present invention;
[0033] Figure 32 is a schematic diagram of the structure of a high-efficiency joint inversion device for large-area airborne gravity and magnetic multi-parameter data according to an embodiment of the present invention;
[0034] Figure 4 is a schematic structural diagram of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION
[0035] The specific implementation modes of the present invention are described in detail below in conjunction with the accompanying drawings, but it should be understood that the protection scope of the present invention is not limited by the specific implementation modes.
[0036] Unless explicitly stated otherwise, throughout the specification and claims, the term “comprise” or variations such as “include” or “comprising”, etc., will be understood to include the stated elements or components but not to exclude other elements or components.
[0037] In order to facilitate the understanding of the embodiments of the present invention, the relevant technologies involved in the embodiments of the present invention are first described below.
[0038] like Figures 1 to 4 As shown, the efficient joint inversion method and device for large-area airborne gravity and magnetic multi-parameter data according to the preferred embodiment of the present invention can be implemented in the following way.
[0039] Figure 1 FIG. 1 is a flow chart of a method for efficiently inverting large-area aerogravimetric and magnetic multi-parameter data according to an embodiment of the present invention. Figure 1 As shown, the method may include the following steps:
[0040] Step 101: adopt a first random algorithm to invert each parameter data in the large-area aerogravimetric and magnetic multi-parameter data to obtain a model disturbance term for each parameter data.
[0041] Specifically, the method for joint inversion of aerogravimetric and magnetic multi-parameter data provided in an embodiment of the present invention is a method suitable for joint inversion of large-area aerogravimetric and magnetic multi-parameter data. Large-area aerogravimetric and magnetic multi-parameter data are data obtained by aerial exploration of a large area. Large-area aerogravimetric and magnetic multi-parameter data include gravity field data and magnetic field data. Gravity field data may include data of one or more parameters (such as gravitational acceleration, etc.). Magnetic field data may also include data of one or more parameters (such as magnetic field intensity, etc.).
[0042] It should be noted that the joint inversion method for airborne gravity and magnetic multi-parameter data provided in an embodiment of the present invention is a joint inversion method for gravity and magnetic multi-parameter data based on a sequential approach. In a sequential approach, the joint inversion of gravity and magnetic multi-parameter data can generally include: independent inversion and minimization of cross-gradient terms. Independent inversion refers to a single (independent) inversion (independent inversion) of gravity (magnetic) different parameter data. Minimization of cross-gradient terms refers to the minimization of cross-gradient terms. In each iteration, a single physical property inversion of different parameter data is first performed to obtain the model disturbance term, and then a cross-gradient term is introduced into the model disturbance term to minimize it, so as to construct a structural coupling relationship between different models and obtain a model correction term; finally, the model is updated according to the model correction term.
[0043] Compared with the independent inversion in the related art, the embodiment of the present invention adopts an efficient strategy for independent inversion. Specifically, other methods with higher efficiency can be used in independent inversion. The method used for independent inversion in the embodiment of the present invention can be called a first random algorithm. The first random algorithm is an efficient random algorithm.
[0044] In some feasible embodiments, a first random algorithm is used to invert each parameter data in the large-area aerogravimetric and magnetic multi-parameter data to obtain a model disturbance term of each parameter data, including:
[0045] For each parameter data, transforming the inversion objective function of the parameter data from a function of a model vector of the parameter data to a first function of a first vector;
[0046] Decomposing the first matrix based on a random singular value decomposition algorithm to obtain a first vector;
[0047] Based on the first vector, the first matrix and the model weighting matrix of the parameter data, a model disturbance term of the parameter data is obtained.
[0048] Specifically, taking any parameter (parameter 1) in the large-area aerogravimetric and magnetic multi-parameter data as an example, the inversion objective function of the parameter data (parameter 1 data) of this parameter is:
[0049]
[0050] Among them, W1 d is the data weighting matrix; G1∈R m×n is the kernel matrix; m1∈R n is the model vector of parameter data; d1 obs ∈R m is the data vector; W1 m =Z1·W1 L1 is the model weight matrix; Z1 and W1 L1 are depth weighted matrix and focus constraint matrix respectively; m1apr is the initial model (referring to the initial state of the model of parameter data); λ1 is the regularization parameter; m and n are the number of data and underground grid divisions respectively. In each iteration of the sequential method, through mathematical transformation, the following formula can be obtained:
[0051]
[0052] Among them, the matrix A1∈R m×n is the first matrix; vector y1 is the first vector; vector d1 is the third vector.
[0053] Based on formula (2), formula (1) can be transformed into the following form:
[0054]
[0055] In the kth iteration of the sequential method, formula (3) is minimized, and we have
[0056] y1 (k) =((A1 (k) ) T A1 (k) +(λ1 (k) ) 2 I n ) -1 (A1 (k) ) T ·d1 (k) (4).
[0057] Among them, I n is the unit matrix. In each embodiment of the present invention, the superscript (k) indicates the kth iteration.
[0058] At this time, the model disturbance term of the parameter data for
[0059]
[0060] In the related art, singular value decomposition can be used to solve formula (4), and its computational cost is O(m 2 ·n), which is more expensive and thus computationally inefficient.
[0061] In the embodiment of the present invention, a random singular value decomposition algorithm may be used to replace the singular value decomposition algorithm. For the matrix A1, its random singular value decomposition is:
[0062]
[0063] Among them, U1 q1 ∈R m×q1 With V1 q1 ∈R n×q1are all orthogonal matrices, which can be represented by the left singular vector u1 i and the right singular vector v1 i Composition; S1 q1 ∈R q1×q1 is the singular value s1 i The diagonal matrix is composed of; the parameter q1 is the target matrix rank parameter, and q1< <m。
[0064] According to the random singular value decomposition result of the first matrix A1, formula (4) can be calculated as follows:
[0065]
[0066] Therefore, based on the random singular value decomposition algorithm, the first matrix A1 is decomposed, and according to formula (7), the first vector y1 can be obtained. After obtaining the first vector y1, the model disturbance term of the parameter data can be obtained according to formula (5):
[0067] Step 102: adopting the target algorithm, based on the model disturbance terms of different parameter data, minimize the cross gradient term processing to obtain the model correction term of each parameter data.
[0068] Specifically, compared with the cross gradient term minimization in the related art, the embodiment of the present invention adopts an efficient strategy to minimize the cross gradient term. Specifically, other more efficient methods can be used in the cross gradient term minimization. The method used for independent inversion in the implementation of the present invention can be called a target algorithm.
[0069] In some feasible embodiments, a target algorithm is used to minimize cross gradient terms based on model disturbance terms of different parameter data to obtain model correction terms of each parameter data, including:
[0070] For different parameter data, the formula for minimizing the cross gradient term is converted into the minimum square norm form;
[0071] Based on the formula for minimizing the cross-gradient term in the form of the least squares norm, the model correction terms for different parameter data are obtained.
[0072] Specifically, during the execution of the target algorithm, the cross-gradient term minimization formula can be re-derived in the form of the least squares norm using matrix transformation to reduce the scale of the original inverse problem.
[0073] Taking two parameter data in the aerogravimetric and magnetic multi-parameter data as an example, in the cross-gradient term minimization part, the cross-gradient term can be minimized by the following formula:
[0074] A (k) Δm (k) =b (k) (8).
[0075] in,
[0076] in, and The model disturbance term obtained by independent inversion of two different parameter data (parameter 1 data and parameter 2 data); Δm1∈R n and Δm2∈R n is the model correction term used to update the model; t∈R 3n is the cross gradient vector; B1∈R 3n×n and B2∈R 3n×n are the partial derivative matrices of t with respect to models m1 and m2 (the model of parameter 1 data and the model of parameter 2 data), respectively; w1, w2, and wt are The weight factor of t. Models m1 and m2 partially construct the structural coupling relationship by minimizing the cross gradient term.
[0077] A∈R 5n×2n The singular value decomposition of can be used to solve formula (8), and its computational cost is O(20n 3 ), which is extremely expensive and thus computationally inefficient.
[0078] In order to achieve efficient calculation, we can re-derive and transform formula (8) into the minimum square norm form:
[0079]
[0080] Formula (10) can be equivalent to
[0081]
[0082] in,
[0083]
[0084] Transforming formula (12), we have:
[0085]
[0086] Based on formula (13), formula (11) can be transformed into
[0087]
[0088] The solution of formula (14) is:
[0089]
[0090] Optionally, the second matrix The singular value decomposition of is used to calculate h. Based on h and the above formula (15), the model correction terms of the two parameter data can be obtained: and
[0091] Using the second matrix The singular value decomposition of is used to calculate h, and the computational cost can be reduced to O(12n 3 ), compared to O(20n 3 ) is reduced to a certain extent, thereby improving the calculation efficiency.
[0092] In some feasible embodiments, the model correction terms for different parameter data are obtained based on the formula for minimizing the cross gradient term in the form of the least squares norm, including:
[0093] Based on the formula of minimizing the cross-gradient term in the form of the least square norm, an alternating direction strategy is used to deal with the cross-gradient term;
[0094] Based on the cross-gradient terms processed by the alternating direction strategy, the formula for minimizing the cross-gradient terms in the form of the least square norm is solved to obtain the model correction terms for different parameter data.
[0095] Specifically, using The singular value decomposition of is used to calculate h, and the computational cost is reduced to a certain extent, but it is still large. In order to further reduce the computational cost and improve the computational efficiency, the alternating direction strategy can be used to process the cross-gradient term under the new formula to further reduce the scale of the inverse problem in the cross-gradient term minimization part. The new formula may include at least one of formulas (10), (11) and (14).
[0096] Considering:
[0097]
[0098] Among them, B1 x ∈R n×n 、B1 y ∈R n×n 、B1 z ∈R n×n are the X, Y, and Z components of B1; B2 x ∈R n×2n 、B2 y ∈R n×2n 、B2 z ∈R n×2n They are the components of B2 in the X, Y, and Z directions respectively.
[0099] An alternating direction strategy can be used to consider only one direction when minimizing the cross-gradient term in each iteration, thereby reducing the size of the inverse problem in the cross-gradient term minimization part compared to considering three directions.
[0100] It should be noted that the model data shown in the above formula (16) is a case where the model data is three-dimensional data (including components in three directions in space), but those skilled in the art can understand the case where the model data is two-dimensional data (including components in two directions in a plane) based on the above embodiment.
[0101] In some feasible embodiments, based on the formula for minimizing the cross-gradient term in the form of the least squares norm, an alternating direction strategy is used to process the cross-gradient term, including:
[0102] A direction is selected as the target direction in turn according to a preset order, and in minimizing the cross gradient term in each iteration, the cross gradient term is processed based on the target direction.
[0103] Specifically, selecting a direction as the target direction in sequence according to a preset order may include a method in which the number of iterations is modulo the number of components included in the model data.
[0104] Optionally, when the model data is three-dimensional data, the current number of iterations k can be divided modulo 3 to determine one of the X, Y, and Z directions as the target direction; or, when the model data is two-dimensional data, the current number of iterations k can be divided modulo 2 to determine one of the X and Y directions as the target direction.
[0105] When the model data is three-dimensional data, there can be:
[0106]
[0107] At this time, the second matrix The dimension is reduced to n×2n, and the computational cost is further reduced to O(2n 3 ), so that the computational efficiency can be further greatly improved.
[0108] In some feasible embodiments, based on the cross-gradient term processed by the alternating direction strategy, the formula for minimizing the cross-gradient term in the form of the least square norm is solved to obtain the model correction terms of the two parameter data, including:
[0109] Decomposing the second matrix based on a random singular value decomposition algorithm to obtain a second vector;
[0110] Based on the second vector and the model disturbance term of the different parameter data, the model correction term of the different parameter data is obtained.
[0111] Specifically, based on the above steps, an efficient random algorithm can be used to minimize the cross-gradient term, thereby further reducing the computational cost of matrix inversion in minimizing the cross-gradient term.
[0112] Second Matrix After the dimension of is reduced to n×2n, the randomized singular value decomposition algorithm can be used instead of the singular value decomposition algorithm to calculate formula (15), minimize the cross-gradient term, and further reduce the computational cost of this part.
[0113] For the second matrix Its random singular value decomposition:
[0114]
[0115] Among them, U q ∈R n×q With V q ∈R 2n×q are all orthogonal matrices, which can be represented by the left singular vector u i and the right singular vector v i Composition; S q ∈R q×q is the singular value s i The diagonal matrix is composed of; q is the target matrix rank parameter, and q< <n。
[0116] Therefore, h in formula (15) can be obtained by the following formula:
[0117]
[0118] At this point, the computational cost of minimizing the cross gradient term is reduced to only O(2·q·n 2 )(q< <n)。
[0119] Step 103: based on the model correction item of each parameter data, update the model of each parameter data until the iterative stop condition is met.
[0120] Specifically, for each parameter data, the model of the parameter data can be updated according to the model correction item of the parameter data until the iterative stop condition is met, thereby obtaining the final result of the inversion.
[0121] The stopping condition for the iteration may be that the number of iterations reaches a preset value or the model of each parameter data has converged.
[0122] Figure 2 This is a second flow chart of an efficient joint inversion method for large-area aerogravimetric and magnetic multi-parameter data according to an embodiment of the present invention. Figure 2 , taking two types of parameter data as an example, the implementation process of the efficient joint inversion method for large-area aerogravimetric and magnetic multi-parameter data is explained.
[0123] The initial models of data 1 (the first parameter data) and data 2 (the second parameter data) are:
[0124] m1apr (1) =m1 ref , m2 apr (1) =m2 ref .
[0125] At this time, the number of iterations k=1.
[0126] Based on the initial model of data 1 and data 2, the number of iterations k satisfies k≤k max In the case of k, independent and efficient inversion steps for data 1 and data 2 can be performed. max is the maximum number of iterations.
[0127] The efficient strategy that can be adopted in this step includes adopting an efficient random algorithm (i.e. the first random algorithm mentioned above). and Thus, the model disturbance term is obtained.
[0128] After the independent efficient inversion steps of data 1 and data 2, an efficient minimization of cross gradient terms step can be performed to minimize the cross gradient terms.
[0129] The efficient strategies that can be used in this step include: deriving new formulas, alternating direction strategies, and efficient random algorithms. This results in the model correction term.
[0130] After minimizing the cross gradient term efficiently, the model can be updated. The specific formula is as follows:
[0131]
[0132] After the model is updated, it can be determined whether the iteration exit condition is met, that is, whether the stop condition of the previous iteration is met.
[0133] If it is satisfied, the final result of the joint inversion can be obtained; if it is not satisfied, the next iteration can be performed, setting k = k + 1, m1 apr (k-1) =m1 (k) , m2 apr (k-1) =m2 (k) , re-execute the independent efficient inversion steps for data 1 and data 2.
[0134] Although the above uses two types of parameter data as an example to illustrate the implementation process of the efficient joint inversion method for large-area airborne gravity and magnetic multi-parameter data, those skilled in the art can generalize it to situations with more than two types of parameter data based on this example and the aforementioned embodiments.
[0135] The beneficial effect of the present invention is that by adopting efficient strategies in both independent inversion and cross-gradient term minimization, efficient calculation of joint inversion of airborne gravity and magnetic multi-parameter data in a sequential manner can be achieved, the efficiency of joint inversion of large-area airborne gravity and magnetic multi-parameter data can be improved, and the application of joint inversion of airborne gravity and magnetic multi-parameter data in large-area actual exploration can be realized.
[0136] The following is a description of the efficient joint inversion device for large-area airborne gravity and magnetic multi-parameter data provided by the present invention. The efficient joint inversion device for large-area airborne gravity and magnetic multi-parameter data described below and the efficient joint inversion method for large-area airborne gravity and magnetic multi-parameter data described above can be referenced to each other.
[0137] Figure 3 Schematic diagram of the structure of the efficient joint inversion device for large-area airborne gravity and magnetic multi-parameter data provided by the present invention. Based on the content of any of the above embodiments, Figure 3 As shown, the device includes an independent inversion module 301, a cross gradient term minimization module 302 and an updating module 303, wherein:
[0138] The independent inversion module 301 is used to use the first random algorithm to invert each parameter data in the large-area aerogravimetric and magnetic multi-parameter data, and obtain the model disturbance term of each parameter data;
[0139] A cross gradient term minimization module 302 is used to adopt a target algorithm to minimize the cross gradient term based on the model disturbance term of different parameter data to obtain the model correction term of each parameter data;
[0140] The updating module 303 is used to update the model of each parameter data based on the model correction item of each parameter data until the iterative stopping condition is met.
[0141] The efficient joint inversion device for large-area airborne gravity and magnetic multi-parameter data provided in an embodiment of the present invention is used to execute the efficient joint inversion method for large-area airborne gravity and magnetic multi-parameter data mentioned above in the present invention. Its implementation method is consistent with the implementation method of the efficient joint inversion method for large-area airborne gravity and magnetic multi-parameter data provided by the present invention, and can achieve the same beneficial effects, which will not be repeated here.
[0142] The efficient joint inversion device for large-area airborne gravity and magnetic multi-parameter data is used for the efficient joint inversion method for large-area airborne gravity and magnetic multi-parameter data in the aforementioned embodiments. Therefore, the description and definition in the efficient joint inversion method for large-area airborne gravity and magnetic multi-parameter data in the aforementioned embodiments can be used for understanding each execution module in the embodiments of the present invention.
[0143] In some feasible implementations, the independent inversion module 301 may be specifically used for:
[0144] For each parameter data, transforming the inversion objective function of the parameter data from a function of a model vector of the parameter data to a first function of a first vector;
[0145] Decomposing the first matrix based on a random singular value decomposition algorithm to obtain a first vector;
[0146] Based on the first vector, the first matrix and the model weighting matrix of the parameter data, a model disturbance term of the parameter data is obtained.
[0147] In some feasible implementations, the cross gradient term minimization module 302 may include:
[0148] A conversion unit, used for converting the formula for minimizing the cross gradient term into a minimum two-norm form for different parameter data;
[0149] The acquisition unit obtains model correction terms for different parameter data based on a formula for minimizing cross gradient terms in the form of a least squares norm.
[0150] In some feasible implementations, the acquisition unit may include:
[0151] A direction alternating subunit is used to minimize the cross-gradient term based on the formula of the least squares norm form, and adopts an alternating direction strategy to process the cross-gradient term;
[0152] The solving subunit is used to solve the formula for minimizing the cross-gradient term in the form of the least square norm based on the cross-gradient term processed by the alternating direction strategy, and obtain the model correction term of different parameter data.
[0153] In some feasible implementations, the direction alternating subunit may be specifically configured to select a direction as a target direction in sequence according to a preset order, and process the cross gradient term based on the target direction in minimizing the cross gradient term in each iteration.
[0154] In some feasible implementations, the solving subunit may be specifically used for:
[0155] Decomposing the second matrix based on a random singular value decomposition algorithm to obtain a second vector;
[0156] Based on the second vector and the model disturbance term of the different parameter data, the model correction term of the different parameter data is obtained.
[0157] The beneficial effect of the present invention is that by adopting efficient strategies in both independent inversion and cross-gradient term minimization, efficient calculation of joint inversion of airborne gravity and magnetic multi-parameter data in a sequential manner can be achieved, the efficiency of joint inversion of large-area airborne gravity and magnetic multi-parameter data can be improved, and the application of joint inversion of airborne gravity and magnetic multi-parameter data in large-area actual exploration can be realized.
[0158] Figure 4 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 4 As shown, the electronic device may include: a processor 410, a communication interface 420, a memory 430 and a communication bus 440, wherein the processor 410, the communication interface 420 and the memory 430 communicate with each other through the communication bus 440. The processor 410 may call the logic instructions in the memory 430 to execute an efficient joint inversion method for large-area aerogravimetric and magnetic multi-parameter data, the method comprising: using a first random algorithm to invert each parameter data in the large-area aerogravimetric and magnetic multi-parameter data, respectively, to obtain a model disturbance term for each parameter data; using a target algorithm to minimize cross-gradient terms based on the model disturbance terms of different parameter data, to obtain a model correction term for each parameter data; based on the model correction term for each parameter data, updating the model of each parameter data until the iterative stop condition is met.
[0159] In addition, the logic instructions in the above-mentioned memory 430 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.
[0160] The processor 410 in the electronic device provided in an embodiment of the present invention can call the logic instructions in the memory 430. Its implementation method is consistent with the implementation method of the efficient joint inversion method of large-area airborne gravity and magnetic multi-parameter data provided by the present invention, and can achieve the same beneficial effects, which will not be repeated here.
[0161] On the other hand, the present invention also provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the efficient joint inversion method of large-area airborne gravity and magnetic multi-parameter data provided by the above-mentioned methods. The method includes: using a first random algorithm to invert each parameter data in the large-area airborne gravity and magnetic multi-parameter data respectively, and obtain a model disturbance term for each parameter data; using a target algorithm to minimize cross-gradient terms based on the model disturbance terms of different parameter data, and obtain a model correction term for each parameter data; based on the model correction term of each parameter data, updating the model of each parameter data until the iteration stop condition is met.
[0162] When the computer program product provided by the embodiment of the present invention is executed, the above-mentioned efficient joint inversion method of large-area airborne gravity and magnetic multi-parameter data is implemented. Its specific implementation method is consistent with the implementation method recorded in the embodiment of the aforementioned method, and can achieve the same beneficial effects, which will not be repeated here.
[0163] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the above-mentioned efficient joint inversion method for large-area airborne gravity and magnetic multi-parameter data, the method comprising: using a first random algorithm to invert each parameter data in the large-area airborne gravity and magnetic multi-parameter data respectively, and obtaining a model disturbance term for each parameter data; using a target algorithm to minimize cross-gradient terms based on the model disturbance terms of different parameter data, and obtaining a model correction term for each parameter data; based on the model correction term of each parameter data, updating the model of each parameter data until the iterative stop condition is met.
[0164] When the computer program stored on the non-transitory computer-readable storage medium provided by the embodiment of the present invention is executed, the above-mentioned efficient joint inversion method of large-area airborne gravity and magnetic multi-parameter data is implemented. Its specific implementation method is consistent with the implementation method recorded in the embodiment of the aforementioned method, and can achieve the same beneficial effects, which will not be repeated here.
[0165] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0166] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0167] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0168] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0169] The foregoing description of specific exemplary embodiments of the present invention is for the purpose of illustration and demonstration. These descriptions are not intended to limit the present invention to the precise form disclosed, and it is clear that many changes and variations can be made based on the above teachings. The purpose of selecting and describing the exemplary embodiments is to explain the specific principles of the present invention and its practical application, so that those skilled in the art can realize and utilize various different exemplary embodiments of the present invention and various different selections and changes. The scope of the present invention is intended to be limited by the claims and their equivalents.
Claims
1. An efficient joint inversion method for large-area aerogravimetric and magnetic multi-parameter data, characterized in that: include: Using a first random algorithm, each parameter data in the large-area aerogravimetric and magnetic multi-parameter data is inverted respectively to obtain a model disturbance term of each parameter data; Adopting the target algorithm, based on the model disturbance terms of different parameter data, the cross gradient term is minimized to obtain the model correction term of each parameter data; Based on the model correction term of each parameter data, the model of each parameter data is updated until the iterative stopping condition is met.
2. The efficient joint inversion method for large-area aerogravimetric and magnetic multi-parameter data according to claim 1 is characterized in that: The first random algorithm is used to invert each parameter data in the large-area aerogravimetric and magnetic multi-parameter data to obtain the model disturbance term of each parameter data, including: For each of the parameter data, transforming the inversion objective function of the parameter data from a function of a model vector of the parameter data to a first function of a first vector; Decomposing the first matrix based on a random singular value decomposition algorithm to obtain the first vector; Based on the first vector and the model weighting matrix of the parameter data, a model disturbance term of the parameter data is obtained.
3. The efficient joint inversion method for large-area aerogravimetric and magnetic multi-parameter data according to claim 1 or 2, characterized in that: The target algorithm is used to minimize the cross gradient term based on the model disturbance term of different parameter data to obtain the model correction term of each parameter data, including: For different parameter data, converting the formula for minimizing the cross gradient term into a minimum two-norm form; Based on the formula for minimizing the cross gradient term in the form of a least squares norm, the model correction term for the different parameter data is obtained.
4. The efficient joint inversion method for large-area aerogravimetric and magnetic multi-parameter data according to claim 3 is characterized in that: The formula for minimizing the cross gradient term based on the least squares norm form, obtaining the model correction term for the different parameter data, comprises: Based on the formula for minimizing the cross-gradient term in the form of the least squares norm, an alternating direction strategy is adopted to process the cross-gradient term; Based on the cross-gradient term processed by the alternating direction strategy, the formula for minimizing the cross-gradient term in the form of the least square norm is solved to obtain the model correction term of the different parameter data.
5. The efficient joint inversion method for large-area aerogravimetric and magnetic multi-parameter data according to claim 4 is characterized in that: The formula for minimizing the cross gradient term based on the least squares norm form uses an alternating direction strategy to process the cross gradient term, including: A direction is selected as a target direction in sequence according to a preset order, and in minimizing the cross gradient term in each iteration, the cross gradient term is processed based on the target direction.
6. The efficient joint inversion method for large-area aerogravimetric and magnetic multi-parameter data according to claim 4 is characterized in that: The cross gradient term processed based on the alternating direction strategy, solving the formula for minimizing the cross gradient term in the form of the least square norm, and obtaining the model correction term for the different parameter data, comprises: Decomposing the second matrix based on a random singular value decomposition algorithm to obtain a second vector; Based on the second vector and the model disturbance term of the different parameter data, a model correction term of the different parameter data is obtained.
7. An efficient joint inversion device for large-area aerogravimetric and magnetic multi-parameter data, characterized in that: include: An independent inversion module is used to adopt a first random algorithm to invert each parameter data in the large-area aerogravimetric and magnetic multi-parameter data, and obtain a model disturbance term of each parameter data; A cross gradient term minimization module is used to adopt a target algorithm to minimize the cross gradient term based on the model disturbance term of different parameter data to obtain the model correction term of each parameter data; An updating module is used to update the model of each parameter data based on the model correction item of each parameter data until the iterative stopping condition is met.
8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the efficient joint inversion method for large-area aerogravimetric and magnetic multi-parameter data as described in any one of claims 1 to 6 are implemented.
9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the efficient joint inversion method for large-area aerogravimetric and magnetic multi-parameter data as claimed in any one of claims 1 to 6 are implemented.
10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the efficient joint inversion method for large-area aerogravimetric and magnetic multi-parameter data as claimed in any one of claims 1 to 6 are implemented.
Citation Information
Patent Citations
Gramian constraint-based gravity-magnetic three-dimensional joint inversion method and system
CN115128700A
Gravity-magnetic three-dimensional joint inversion method based on cross gradient constraint
CN115201927A
A starting model independent full waveform inversion system, method, and computer-program product for subsurface velocity estimation
US20240210587A1
Load identification method and system based on hybrid iterative regularization
WO2023246117A1
