Prestack inversion method and device based on gradient optimizer

Abstract

The present invention provides a prestack inversion method and device based on a gradient optimizer, relating to the technical field of prestack inversion, including: obtaining prestack angle gather data, horizon data, and elastic parameters of a target well log in a target work area; determining a solution space for the elastic parameters in the target work area based on the elastic parameters and horizon data; constructing a first objective function for inverting the elastic parameters in the target work area based on the prestack angle gather data of the target work area and the Zoeppritz equation; determining a first population within the solution space for solving the elastic parameters of a target seismic trace; iteratively updating the first population using a gradient optimizer to determine an optimal solution for the elastic parameters of the target seismic trace based on the angle gather data of the target seismic trace and the first objective function; and constructing an elastic parameter profile of the target work area based on the optimal elastic parameter solutions of multiple seismic traces. The application of the gradient optimizer enables the method of the present invention to simultaneously balance the convergence speed and inversion accuracy of prestack inversion.

Description

Technical Field

[0001] The present invention relates to the technical field of prestack inversion, and in particular to a prestack inversion method and device based on a gradient optimizer. Background Art

[0002] Prestack inversion is a multidimensional optimization problem. Gradient-based optimization algorithms, such as Gauss-Newton and conjugate gradient methods, offer very fast optimization speeds and are the mainstream approach for large-scale inversion. However, they are overly dependent on the initial model. When the initial model lacks sufficient accuracy, the inversion often fails to achieve satisfactory results. Although intelligent optimization algorithms, such as the particle swarm optimization algorithm, combined with various other optimization algorithms, can achieve relatively accurate inversion results without relying on the initial model, they require a very large amount of computation and consume a considerable amount of computing time, limiting their applicability to practical work areas.

[0003] In summary, the existing pre-stack inversion methods have the technical problem of being unable to simultaneously meet the requirements of fast convergence speed and high inversion accuracy. Summary of the Invention

[0004] The object of the present invention is to provide a prestack inversion method and apparatus based on a gradient optimizer, so as to alleviate the technical problem that the existing prestack inversion methods cannot simultaneously meet the requirements of fast convergence speed and high inversion accuracy.

[0005] In a first aspect, the present invention provides a pre-stack inversion method based on a gradient optimizer, comprising: obtaining pre-stack angle gather data, layer data and elastic parameters of a target well logging of a target work area; wherein the pre-stack angle gather data is a collection of angle gather data of multiple seismic channels; the target well logging represents any well logging in the target work area; based on the elastic parameters of the target well logging and the layer data of the target work area, determining the solution space of the elastic parameters in the target work area; based on the pre-stack angle gather data of the target work area and the Zoeppritz equation, constructing a method for inverting the elastic parameters in the target work area. A first objective function; under the constraints of the solution space, randomly initializing a first population for solving the elastic parameters of the target seismic trace; wherein the target seismic trace represents any one of the multiple seismic traces; in the first population, the position of each individual corresponds one-to-one to a potential solution of the elastic parameters of the target seismic trace; using a gradient optimizer to iteratively update the first population to determine the optimal solution of the elastic parameters of the target seismic trace based on the angular trace data of the target seismic trace and the first objective function; and constructing an elastic parameter profile of the target work area based on the optimal solutions of the elastic parameters of the multiple seismic traces.

[0006] In an optional embodiment, based on the pre-stack angle gather data and the Zoeppritz equation of the target work area, a first objective function for inverting the elastic parameters in the target work area is constructed, including: superimposing the pre-stack angle gather data of the target work area according to a specified number of incident angles to obtain post-stack seismic data; extracting the seismic wavelet data of the target work area from the post-stack seismic data; constructing a convolution function from the elastic parameters to the synthetic seismic data based on the seismic wavelet data and the Zoeppritz equation; and constructing the first objective function based on the convolution function, the target regularization parameter, the preset selection matrix and the preset scale matrix.

[0007] In an optional embodiment, the method further includes: constructing a second objective function for inverting the elastic parameters of the target logging based on the pre-stack angle track data of the target logging, the convolution function, the initial regularization parameter, the preset selection matrix and the preset scale matrix; randomly initializing a second population for solving the elastic parameters of the target logging under the constraints of the solution space; wherein, in the second population, the position of each individual corresponds one-to-one to a potential solution of the elastic parameters of the target logging; iteratively updating the second population using a gradient optimizer to determine the initial solution of the elastic parameters of the target seismic trace based on the second objective function; when it is determined that the error between the elastic parameters of the target logging and the initial solution does not meet the preset error requirement, adjusting the initial regularization parameter and returning to the previous step until the error between the initial solution and the elastic parameters of the target logging meets the preset error requirement, thereby obtaining the target regularization parameter.

[0008] In an optional embodiment, the first population is iteratively updated using a gradient optimizer to determine the optimal solution for the elastic parameters of the target seismic trace based on the angle trace data of the target seismic trace and the first objective function, including: determining a third objective function for inverting the elastic parameters of the target seismic trace based on the angle trace data of the target seismic trace and the first objective function; repeating the following steps until a preset number of iterations is reached, and using the global optimal individual position that reaches the preset number of iterations as the elastic parameters of the target seismic trace: determining the current global optimal individual position and the global worst individual position based on the current population and the third objective function; wherein, in the first iteration, the current population is the first population; based on the gradient in the gradient optimizer The degree search rule processes the current position of the target individual, the current position of the random individual, the global optimal individual position and the global worst individual position to obtain the first displacement vector direction, the second displacement vector direction and the third displacement vector direction of the target individual; wherein the target individual represents any individual in the current population; based on the first displacement vector direction, the second displacement vector direction and the third displacement vector direction of the target individual, the planned update position of the target individual is determined; based on the local escape operator in the gradient optimizer, the planned update position, the global optimal individual position, the first displacement vector direction, the second displacement vector direction and the current position of the random individual are processed to obtain the position of the target individual in the next generation.

[0009] In an optional embodiment, the first displacement vector direction is expressed as: The second displacement vector direction is expressed as: The third displacement vector direction is expressed as: The planned updated position of the target individual is expressed as: in, Indicates the position of the vth individual in the kth iteration, randn represents a random number in the range [0, 1] that conforms to the normal distribution, ρ1 = 2 × rand × α - α, rand represents a random value in [0, 1], Indicates that β min represents the first preset value, β max represents the second preset value, K represents the preset number of iterations, Δm=rand(1:3f)×|step|, rand(1:3f) represents a random number between [0, 1] in the 3f dimension, f represents the number of sampling points in each seismic trace, r1, r2, r3, r4, v are different from each other, r1, r2, r3, r4 represent four integers randomly selected from [1, N], N represents the total number of individuals in the population, m best represents the global optimal individual position, m worst represents the global worst individual position, ε represents a very small number greater than zero, ρ2=2×rand×α-α, r a and r b Represents two random numbers between [0, 1].

[0010] In an optional embodiment, the position of the target individual in the next generation is expressed as:

[0011]

[0012] Among them, f1 represents a random number in [-1,1], rand and μ1 represent random numbers between [0,1]. μ2 represents a random number between [0,1], m rand =M min +rand×(M max -M min ) indicates that M max represents the maximum boundary of the solution space, M min represents the minimum boundary of the solution space, represents an individual randomly selected from the current population, and f2 represents a random number from a normal distribution with a variance of 0 and a standard deviation of 1.

[0013] In an optional embodiment, the first objective function is expressed as: Among them, d obs Represents the pre-stack angle gather data of the area to be inverted, d cal represents the convolution function, λ represents the target regularization parameter, f represents the number of sampling points in each seismic trace, m represents the elastic parameter to be solved, Φ j =(D j ) T ψ -1 (D j ), D j represents the preset selection matrix, with a dimension of 3*3f, and ψ represents the preset scale matrix, with a dimension of 3*3.

[0014] In a second aspect, the present invention provides a pre-stack inversion device based on a gradient optimizer, an acquisition module for acquiring pre-stack angle gather data, layer data and elastic parameters of a target well logging of a target work area; wherein the pre-stack angle gather data is a collection of angle gather data of multiple seismic channels; the target well logging represents any well logging in the target work area; a determination module for determining the solution space of the elastic parameters in the target work area based on the elastic parameters of the target well logging and the layer data of the target work area; a first construction module for constructing a first target well logging solution for inverting the elastic parameters in the target work area based on the pre-stack angle gather data of the target work area and the Zoeppritz equation. function; an initialization module for randomly initializing a first population for solving the elastic parameters of the target seismic trace under the constraints of the solution space; wherein the target seismic trace represents any seismic trace among the multiple seismic traces; in the first population, the position of each individual corresponds one-to-one to a potential solution of the elastic parameters of the target seismic trace; an update and determination module for iteratively updating the first population using a gradient optimizer to determine the optimal solution of the elastic parameters of the target seismic trace based on the angle trace data of the target seismic trace and the first objective function; a second construction module for constructing an elastic parameter profile of the target work area based on the optimal solution of the elastic parameters of the multiple seismic traces.

[0015] In a third aspect, the present invention provides an electronic device comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and when the processor executes the computer program, the steps of the pre-stack inversion method based on the gradient optimizer described in any one of the aforementioned embodiments are implemented.

[0016] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer instructions, which, when executed by a processor, implement the pre-stack inversion method based on a gradient optimizer as described in any one of the aforementioned embodiments.

[0017] The present invention provides a pre-stack inversion method based on a gradient optimizer. After determining the solution space of elastic parameters in the target work area and the first objective function for inverting the elastic parameters, the method needs to randomly initialize a first population for solving the elastic parameters of the target seismic trace under the constraints of the solution space, and then use the gradient optimizer to iteratively update the first population to determine the optimal solution for the elastic parameters of the target seismic trace based on the angle trace data of the target seismic trace and the first objective function. The gradient search rule GSR of the gradient optimizer can enhance the exploration trend, accelerate the convergence rate, and obtain better individual positions in the search space; the local escape operator LEO can enable the method to escape from the local optimal solution and improve the global search capability of the algorithm. Therefore, the method of the present invention can simultaneously take into account the convergence speed and inversion accuracy of pre-stack inversion. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0019] Figure 1 A flowchart of a prestack inversion method based on a gradient optimizer provided in an embodiment of the present invention;

[0020] Figure 2 A flowchart of determining an optimal solution for elastic parameters of a target seismic trace provided by an embodiment of the present invention;

[0021] Figure 3 A functional module diagram of a prestack inversion device based on a gradient optimizer provided in an embodiment of the present invention;

[0022] Figure 4 A schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0024] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention as claimed, but rather merely represents selected embodiments of the present invention. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without creative effort shall fall within the scope of protection of the present invention.

[0025] The following embodiments of the present invention are described in detail with reference to the accompanying drawings. In the absence of conflict, the following embodiments and features in the embodiments may be combined with each other.

[0026] Example 1

[0027] Figure 1 A flowchart of a prestack inversion method based on a gradient optimizer is provided in an embodiment of the present invention. Figure 1 As shown, the method specifically includes the following steps:

[0028] Step S102: obtaining pre-stack angle gather data, horizon data and elastic parameters of target well logging in the target work area.

[0029] The pre-stack angle gather data is a collection of angle gather data of multiple seismic traces; the target well logging refers to any well logging in the target work area.

[0030] Specifically, first obtain the pre-stack angle gather data with a high signal-to-noise ratio in the target work area after data processing as the actual observation data. The pre-stack angle gather data can be obtained by the following operations: obtain the seismic data of the target work area, process the seismic data, and convert it into an angle gather. This embodiment does not specifically limit the method for obtaining pre-stack angle gather data, and the user can also obtain it by processing other data. In order to accurately invert the elastic parameters in the target work area, it is also necessary to obtain the layer data of the target work area and the elastic parameters of the target well logging. The elastic parameters of the target well logging can be obtained by interpreting the well logging. The elastic parameters in this embodiment include: longitudinal wave velocity v p , shear wave velocity v s and density ρ.

[0031] Step S104 : determining a solution space of the elastic parameters in the target work area based on the elastic parameters of the target well logging and the horizon data of the target work area.

[0032] In an embodiment of the present invention, the elastic parameters of the target well logging data are used as reference data and combined with the horizon data of the target work area to construct an initial model of the elastic parameters within the target work area. Specifically, an appropriate frequency threshold is set to remove high-frequency data and retain low-frequency data. Through interpolation and extrapolation of the elastic parameters of the target well logging data and the horizon data, an initial model of the low-frequency P-wave velocity, S-wave velocity, and density within the target work area is obtained. After obtaining the initial model, the range of values ​​within the initial model, which is within a preset offset value, is used as the solution space for the elastic parameters within the target work area. For example, ±30% of the initial model is used as the search space for the optimization algorithm, i.e., the solution space for the elastic parameters within the target work area.

[0033] Step S106 : constructing a first objective function for inverting elastic parameters in the target work area based on the pre-stack angle gather data of the target work area and the Zoeppritz equation.

[0034] The goal of the prestack inversion in the embodiment of the present invention is to minimize the residual between the synthetic data of the forward convolution based on the inversion result and the actual observation data. The embodiment of the present invention adopts the accurate Zoeppritz equation for forward modeling. According to the Zoeppritz equation, the functional relationship from the elastic parameter to the reflection coefficient sequence can be determined: R pp =R(m,θ), where m represents the elastic parameter, θ represents the incident angle set, and R pprepresents the reflection coefficient sequence. Seismic wavelets can be extracted from prestack angle gather data. Synthetic data can be obtained by convolving the seismic wavelets with the reflection coefficient sequence. Residual terms can be determined based on the synthetic data and actual observation data (i.e., prestack angle gather data). By combining these residual terms with the constraint terms, a first objective function can be constructed for inverting elastic parameters within the target area.

[0035] Step S108: Under the constraints of the solution space, randomly initialize a first population for solving the elastic parameters of the target seismic trace.

[0036] Specifically, to complete the inversion of elastic parameters in the target work area, the embodiment of the present invention adopts a single-channel form. In other words, each seismic channel in the seismic profile is inverted separately, and finally the inversion results of all seismic channels are synthesized to obtain a complete inversion parameter profile of the target work area.

[0037] To solve for the elastic parameters of a target seismic trace, given a fixed solution space, a first population must be randomly initialized within the solution space. The target seismic trace represents any one of multiple seismic traces. Within the first population, the position of each individual corresponds one-to-one to a potential solution for the elastic parameters of the target seismic trace. A larger number of individuals in the population enhances global search capabilities, but also increases computational time. This embodiment of the present invention does not impose a specific limit on the size of the first population; users can set it based on their needs.

[0038] If the number of sampling points of the target seismic trace is f, and the elastic parameters of each sampling point include three parameters [v p v s ρ], then it can be seen that the data dimension of the elastic parameters of the target seismic trace is 3f, that is, the data dimension of each individual in the first population should also be 3f.

[0039] Step S110 , iteratively updating the first population using a gradient optimizer to determine an optimal solution for the elastic parameters of the target seismic trace based on the angle gather data of the target seismic trace and the first objective function.

[0040] The gradient-based optimizer (GBO) utilizes two main operators: the gradient search rule (GSR) and the local escaping operator (LEO). GSR uses a gradient-based approach to enhance exploration trends, accelerate convergence, and achieve a better position in the search space. LEO enables GBO to escape local optima, improving the algorithm's global search capabilities. Therefore, the GBO algorithm does not require a large population like the particle swarm optimization algorithm and can achieve convergence with fewer individuals.

[0041] During the solution process, the angle track data of the target seismic channel is used as the actual observation data in the first objective function, and the individual position is substituted into the first objective function as the solution to calculate the value of the first objective function. The value of the first objective function is continuously reduced through population iteration until the preset convergence condition is reached, for example, the specified convergence accuracy is met, or the specified number of iterations is reached, thereby determining the optimal solution for the elastic parameters of the target seismic channel.

[0042] The application of gradient optimizer can effectively avoid the high computing power consumption and premature convergence problems of particle swarm type methods for high-dimensional optimization problems, better balance the early exploration stage and the later development stage of the metaheuristic algorithm, and accelerate the convergence speed.

[0043] Step S112: constructing an elastic parameter profile of the target work area based on the optimal solutions of the elastic parameters of the multiple seismic traces.

[0044] An embodiment of the present invention provides a pre-stack inversion method based on a gradient optimizer. After determining the solution space of the elastic parameters in the target work area and the first objective function for inverting the elastic parameters, the method needs to randomly initialize a first population for solving the elastic parameters of the target seismic trace under the constraints of the solution space, and then use the gradient optimizer to iteratively update the first population to determine the optimal solution for the elastic parameters of the target seismic trace based on the angle trace data of the target seismic trace and the first objective function. The gradient search rule GSR of the gradient optimizer can enhance the exploration trend, accelerate the convergence rate, and obtain a better individual position in the search space; the local escape operator LEO can enable the method to escape from the local optimal solution and improve the global search capability of the algorithm. Therefore, the method of the present invention can simultaneously take into account the convergence speed and inversion accuracy of pre-stack inversion.

[0045] In an optional embodiment, the above step S106, based on the pre-stack angle gather data of the target work area and the Zoeppritz equation, constructs a first objective function for inverting the elastic parameters in the target work area, which specifically includes the following steps:

[0046] Step S1061 : stacking the pre-stack angle gather data of the target work area according to a specified number of incident angles to obtain post-stack seismic data.

[0047] Step S1062: extracting seismic wavelet data of the target work area from the post-stack seismic data.

[0048] The method of stacking the pre-stack angle gather data of the target work area by angle to obtain post-stack seismic data, and then extracting seismic wavelet data from the post-stack seismic data is a well-known technology in the art and will not be described in detail here.

[0049] Step S1063: Based on the seismic wavelet data and the Zoeppritz equation, a convolution function of the elastic parameters to the synthetic seismic data is constructed.

[0050] To obtain the convolution function, first calculate the reflection coefficient sequence according to the Zoeppritz equation: R pp =R(m,θ), where m = [t1 t2…t f ] T , f represents the total number of sampling points in the target seismic trace, and the elastic parameters (P-wave velocity, S-wave velocity and density) corresponding to the j-th sampling point are expressed as: j =[v p v s ρ] T j ; θ={θ1,θ2,…θ l}, l represents the total number of incident angles, and the reflection coefficient sequence R pp Combined with the seismic wavelet data W(θ) extracted in the previous step, the convolution function of the elastic parameters to the synthetic seismic data can be constructed. That is, the forward convolution process of single-channel synthetic data is as follows:

[0051] in,

[0052] d cal represents the single-channel forward synthetic data, that is, the convolution function.

[0053] Step S1064: construct a first objective function based on the convolution function, the target regularization parameter, the preset selection matrix and the preset scaling matrix.

[0054] In this embodiment of the present invention, the first objective function is expressed as: Among them, d obs Represents the pre-stack angle gather data of the area to be inverted, d cal represents the convolution function, λ represents the target regularization parameter, f represents the number of sampling points in each seismic trace, m represents the elastic parameter to be solved, Φ j =(D j ) T ψ -1 (D j ), D j represents the preset selection matrix with a dimension of 3*3f, and ψ represents the preset scale matrix with a dimension of 3*3.

[0055] According to the first objective function d obs From the parameter definition, we can see that d obsAlthough it is pre-stack angle gather data (i.e., actual observation data), its value needs to be determined according to the actual area to be inverted. In other words, if the elastic parameters of the target seismic trace are to be inverted, then d in the first objective function obs The pre-stack angle gather data of the target seismic trace should be brought in.

[0056] The construction principle of the first objective function is introduced in detail below:

[0057] The goal of prestack inversion is to obs The specific method is to use m convolution to synthesize data, that is, d cal , so that it is consistent with the actual observation data d obs The inversion is performed under the Bayesian framework. Since the elastic parameter m has strong multi-solution characteristics during the inversion process, the prior information should be used as a regularization term to constrain the inversion process. Based on this, we can obtain: Among them, P(m|d obs ) is the posterior probability density distribution function, P(d obs |m) is the likelihood function of the observed data, P(d obs ) is the marginal probability of the observed data, if d obs It is known that the marginal probability of the observed data is a constant value, and P(m) is the prior probability function of the elasticity parameter.

[0058] Assume that the noise follows a zero-mean Gaussian distribution with a variance of Then the likelihood function P(d obs |m) can be expressed as: P(m) provides prior information about the unknown parameters to constrain the inversion. Because elastic parameters have strong statistical correlations, prior information from the target well log is introduced to mitigate uncertainty. Different prior distributions have different characteristics: a Gaussian distribution can produce a smooth solution, while a sparse distribution can produce a sparse solution. This embodiment of the present invention utilizes a three-variable Cauchy distribution as prior information. It adds a scaling matrix describing the correlation between the three parameters to the traditional Cauchy distribution. The prior probability function is shown below: Among them, Φ j =(D j ) T ψ -1 (D j ), Dj represents the preset selection matrix with a dimension of 3*3f, which is used to select from m=[t1 t2…t f ] T Select the parameter vector t at the jth sampling point j; ψ represents a preset scale matrix of dimension 3*3, which can be solved using the maximum expectation algorithm (EM), including finding the expected value (E) and maximization (M).

[0059] Based on the above P(m|d obs ), P(d obs The expressions of P(m) and P(m) can be used to obtain the posterior probability distribution of the elastic parameter m:

[0060] The purpose of the Bayesian inversion framework is to find appropriate parameters to maximize the maximum a posteriori probability. Solving the maximum a posteriori probability formula is equivalent to solving the minimum absolute value of the exponential term. Therefore, the first objective function is set as:

[0061] The m that can minimize the value of the first objective function is the final inversion result. The intelligent optimization algorithm can be used to search for the optimal solution within a given search interval for fitting convergence. The target regularization parameter λ is used to balance the residual term in the first objective function. and constraints The weight between .

[0062] If the weights of the balanced residual term and the constraint term in the first objective function are determined only through experience, if the values ​​are inappropriate, the accuracy of the elastic parameters ultimately obtained will be affected. Therefore, the embodiment of the present invention proposes a method flow for optimizing the regularization parameter by using the pre-stack angle gather data of the target well logging to invert the elastic parameters of the target well logging (elastic parameter prediction values), and using the elastic parameters of the target well logging in step S102 (elastic parameter actual values) as the inversion targets of the inversion results. Specifically, the method of the present invention further includes the following steps:

[0063] Step S201 : constructing a second objective function for inverting elastic parameters of the target well logging based on pre-stack angle gather data of the target well logging, a convolution function, an initial regularization parameter, a preset selection matrix and a preset scaling matrix.

[0064] The first objective function is known to be: Substitute the pre-stack angle gather data of the target well into the first objective function to obtain the second objective function, which is expressed as: Where d′ obs represents the pre-stack angle gather data of the target logging, and λ0 represents the initial regularization parameter.

[0065] Step S202 : randomly initializing a second population for solving the elastic parameters of the target well logging under the constraints of the solution space; wherein, in the second population, the position of each individual corresponds one-to-one to a potential solution of the elastic parameters of the target well logging.

[0066] Step S203: Iteratively update the second population using a gradient optimizer to determine an initial solution of elastic parameters of the target seismic trace based on the second objective function.

[0067] In the second population, the position of each individual represents a solution of the second objective function. Substituting the solution into the second objective function, we can get the corresponding second objective function value, that is, the fitness function value of the individual. Based on the construction principle of the first objective function above, we can know that the smaller the value of the second objective function, the higher the posterior probability P(m|d obs ), the better the position of the corresponding individual. In view of this, when the gradient optimizer is used to iteratively update the second population, the position of the individual is continuously changed to make the value of the second objective function smaller and smaller until convergence. The individual position that minimizes the current value of the second objective function is then used as the initial solution of the elastic parameters of the target seismic trace.

[0068] In step S204, if it is determined that the error between the elastic parameter of the target well logging and the initial solution does not meet the preset error requirement, the initial regularization parameter is adjusted, and the process returns to the previous step until the error between the initial solution and the elastic parameter of the target well logging meets the preset error requirement, thereby obtaining the target regularization parameter.

[0069] If the initial solution obtained by optimizing the second objective function using the initial regularization parameter has a large error with the actual elastic parameters of the target well logging, that is, it cannot meet the preset error requirements, then it means that the value of the initial regularization parameter is inappropriate. In the next step, the value of the initial regularization parameter should be adjusted to update the second objective function. Then, based on the updated second objective function, the gradient optimizer is used to iteratively update the second population again, and the individual position that can minimize the updated second objective function value (that is, the updated initial solution) is sought until the error between the updated initial solution and the elastic parameters (actual values) of the target well logging meet the preset error requirements. At this time, the adjustment of the regularization parameter value can be stopped to obtain the target regularization parameter.

[0070] In an optional embodiment, referring to Figure 2 The above step S110, using the gradient optimizer to iteratively update the first population to determine the optimal solution of the elastic parameters of the target seismic trace based on the angle gather data of the target seismic trace and the first objective function, specifically includes the following steps:

[0071] Step S1101 : determining a third objective function for inverting elastic parameters of the target seismic trace based on the angle gather data of the target seismic trace and the first objective function.

[0072] Specifically, the d in the first objective function is obsBy setting the angle gather data of the target seismic trace as the target seismic trace, the third objective function for inverting the elastic parameters of the target seismic trace can be obtained.

[0073] Repeat steps S1102 to S1105 until a preset number of iterations is reached, and use the globally optimal individual position that reaches the preset number of iterations as the elastic parameter of the target seismic trace:

[0074] Step S1102: Determine the current global optimal individual position and the global worst individual position based on the current population and the third objective function.

[0075] Under the definition of the third objective function, the global optimal individual position represents the individual position that minimizes the third objective function value up to the current point in time; the global worst individual position represents the individual position that maximizes the third objective function value up to the current point in time. Obviously, in order to determine the global optimal and worst individual positions, it is necessary to calculate the third objective function value for each individual in the current population.

[0076] Among them, in the first iteration, the current population is the first population.

[0077] Step S1103, based on the gradient search rule in the gradient optimizer, the current position of the target individual, the current position of the random individual, the global optimal individual position and the global worst individual position are processed to obtain the first displacement vector direction, the second displacement vector direction and the third displacement vector direction of the target individual.

[0078] The target individual represents any individual in the current population.

[0079] Step S1104: determining a planned update position of the target individual based on the first displacement vector direction, the second displacement vector direction, and the third displacement vector direction of the target individual.

[0080] The gradient search rule GSR helps the algorithm consider random behavior in the optimization process and promotes global exploration. The direction of movement (DM) is used to create appropriate local search trends to improve the convergence speed of the GBO algorithm. Based on GSR and DM, in this embodiment of the present invention, the first displacement vector direction is expressed as:

[0081]

[0082] The second displacement vector direction is expressed as:

[0083]

[0084] The direction of the third displacement vector is expressed as:

[0085] The planned updated position of the target individual is expressed as:

[0086]

[0087] in, Indicates the position of the vth individual in the kth iteration, randn represents a random number in the range [0, 1] that conforms to the normal distribution, ρ1 = 2 × rand × α - α, rand represents a random value in [0, 1], Indicates that β min represents the first preset value, β max represents the second preset value, K represents the preset number of iterations, Δm = rand(1:3f)×|step|, rand(1:3f) represents a random number between [0, 1] in the 3f dimension, and f represents the number of sampling points in each seismic trace. r1, r2, r3, r4, v are different from each other, r1, r2, r3, r4 represent four integers randomly selected from [1, N], N represents the total number of individuals in the population, m best represents the global optimal individual position, m worst represents the global worst individual position, ε represents a very small number greater than zero, ρ2=2×rand×α-α, r a and r b Represents two random numbers between [0, 1].

[0088] Step S1105, based on the local escape operator in the gradient optimizer, the planned update position, the global optimal individual position, the first displacement vector direction, the second displacement vector direction and the current position of the random individual are processed to obtain the position of the target individual in the next generation.

[0089] The algorithm has a strong exploration capability in the early stage to obtain suitable particles, and a high development capability in the later stage to accelerate the convergence speed. The local escape operator LEO generates a position with superior performance through multiple solutions.

[0090] In an optional embodiment, the position of the target individual in the next generation is expressed as:

[0091]

[0092] Among them, f1 represents a random number in [-1,1], rand and μ1 represent random numbers between [0,1]. μ2 represents a random number between [0,1], m rand =M min +rand×(M max -Mmin ) indicates that M max represents the maximum boundary of the solution space, M min Indicates the minimum boundary of the solution space. If the initial model ±30% is the solution space, then M max and M min The values ​​of are: initial model +30% position and -30% position; represents a randomly selected individual in the current population, and f2 represents a random number from a normal distribution with a variance of 0 and a standard deviation of 1.

[0093] By repeatedly executing steps S1102 to S1105, the positions of individuals in the population in the next generation are continuously updated, so that the value of the third objective function is continuously reduced until the preset number of iterations is reached (the convergence accuracy is met by default), then the iteration is stopped, and the global optimal individual position when the iteration is stopped is used as the elastic parameter of the target seismic trace.

[0094] It has been verified that the gradient oscillator optimizer (GBO) algorithm is compared with the particle swarm optimization (PSO) algorithm. When achieving similar minimum fitness values, the GBO algorithm requires fewer iterations and individuals, takes less time, and is more efficient.

[0095] In summary, the use of the gradient-based optimizer (GBO) algorithm effectively avoids the problem of premature convergence and local minima during the optimization process. The inclusion of the gradient search rule allows for faster convergence, greater stability, and greater suitability for high-dimensional search problems. It can globally search a given solution space, yielding more accurate inversion results. Furthermore, the entire inversion process is more stable, with significantly enhanced noise immunity, enabling clearer resolution of geological structures. The inverted synthetic data is more consistent with real data, effectively improving the resolution and accuracy of prestack inversion.

[0096] Example 2

[0097] An embodiment of the present invention also provides a pre-stack inversion device based on a gradient optimizer, which is mainly used to execute the pre-stack inversion method based on a gradient optimizer provided in the above-mentioned embodiment 1. The following is a detailed introduction to the pre-stack inversion device based on a gradient optimizer provided in an embodiment of the present invention.

[0098] Figure 3 is a functional module diagram of a prestack inversion device based on a gradient optimizer provided by an embodiment of the present invention, such as Figure 3 As shown, the device mainly includes: an acquisition module 11, a determination module 12, a first construction module 13, an initialization module 14, an update and determination module 15, and a second construction module 16, wherein:

[0099] The acquisition module 11 is used to obtain pre-stack angle gather data, layer data and elastic parameters of target well logging in the target work area; wherein the pre-stack angle gather data is a collection of angle gather data of multiple seismic channels; the target well logging represents any well logging in the target work area.

[0100] The determination module 12 is used to determine the solution space of the elastic parameters in the target work area based on the elastic parameters of the target well logging and the layer data of the target work area.

[0101] The first construction module 13 is used to construct a first objective function for inverting elastic parameters in the target work area based on the pre-stack angle gather data of the target work area and the Zoeppritz equation.

[0102] The initialization module 14 is used to randomly initialize a first population for solving the elastic parameters of the target seismic trace under the constraints of the solution space; wherein the target seismic trace represents any seismic trace among multiple seismic traces; in the first population, the position of each individual corresponds one-to-one to a potential solution of the elastic parameters of the target seismic trace.

[0103] The updating and determining module 15 is configured to iteratively update the first population using a gradient optimizer to determine an optimal solution for the elastic parameters of the target seismic trace based on the angle gather data of the target seismic trace and the first objective function.

[0104] The second construction module 16 is used to construct an elastic parameter profile of the target work area based on the optimal solution of elastic parameters of multiple seismic traces.

[0105] An embodiment of the present invention provides a pre-stack inversion device based on a gradient optimizer. After determining the solution space of the elastic parameters in the target work area and the first objective function for inverting the elastic parameters, the device needs to randomly initialize a first population for solving the elastic parameters of the target seismic trace under the constraints of the solution space, and then use the gradient optimizer to iteratively update the first population to determine the optimal solution for the elastic parameters of the target seismic trace based on the angle trace data of the target seismic trace and the first objective function. The gradient search rule GSR of the gradient optimizer can enhance the exploration trend, accelerate the convergence rate, and obtain a better individual position in the search space; the local escape operator LEO can enable the method to escape from the local optimal solution and improve the global search capability of the algorithm. Therefore, the device of the present invention can simultaneously take into account the convergence speed and inversion accuracy of pre-stack inversion.

[0106] Optionally, the first building module 13 is specifically configured to:

[0107] The pre-stack angle gather data of the target area are stacked according to a specified number of incident angles to obtain post-stack seismic data.

[0108] Extract seismic wavelet data of the target area from post-stack seismic data.

[0109] Based on seismic wavelet data and Zoeppritz equation, a convolution function from elastic parameters to synthetic seismic data is constructed.

[0110] A first objective function is constructed based on the convolution function, the target regularization parameter, the preset selection matrix and the preset scaling matrix.

[0111] Optionally, the device is further used to:

[0112] Based on the pre-stack angle gather data of the target well logging, the convolution function, the initial regularization parameter, the preset selection matrix and the preset scale matrix, a second objective function for inverting the elastic parameters of the target well logging is constructed.

[0113] Under the constraints of the solution space, a second population for solving the elastic parameters of the target well logging is randomly initialized; wherein, in the second population, the position of each individual corresponds one-to-one to a potential solution of the elastic parameters of the target well logging.

[0114] The second population is iteratively updated using a gradient optimizer to determine an initial solution of elastic parameters of the target seismic trace based on a second objective function.

[0115] When it is determined that the error between the elastic parameter of the target well logging and the initial solution does not meet the preset error requirement, the initial regularization parameter is adjusted and the previous step is returned to be executed until the error between the initial solution and the elastic parameter of the target well logging meets the preset error requirement, thereby obtaining the target regularization parameter.

[0116] Optionally, the updating and determining module 15 is specifically configured to:

[0117] A third objective function for inverting elastic parameters of the target seismic trace is determined based on the angle gather data of the target seismic trace and the first objective function.

[0118] Repeat the following steps until the preset number of iterations is reached, and use the globally optimal individual position that reaches the preset number of iterations as the elastic parameter of the target seismic trace:

[0119] Based on the current population and the third objective function, the current global optimal individual position and the global worst individual position are determined; wherein, in the first iteration, the current population is the first population.

[0120] Based on the gradient search rule in the gradient optimizer, the current position of the target individual, the current position of the random individual, the global optimal individual position and the global worst individual position are processed to obtain the first displacement vector direction, the second displacement vector direction and the third displacement vector direction of the target individual; wherein the target individual represents any individual in the current population.

[0121] A planned updated position of the target individual is determined based on the first displacement vector direction, the second displacement vector direction, and the third displacement vector direction of the target individual.

[0122] Based on the local escape operator in the gradient optimizer, the planned update position, the global optimal individual position, the first displacement vector direction, the second displacement vector direction and the current position of the random individual are processed to obtain the position of the target individual in the next generation.

[0123] Optionally, the first displacement vector direction is expressed as: The second displacement vector direction is expressed as: The direction of the third displacement vector is expressed as: The planned updated position of the target individual is expressed as: in, Indicates the position of the vth individual in the kth iteration, randn represents a random number in the range [0, 1] that conforms to the normal distribution, ρ1 = 2 × rand × α - α, rand represents a random value in [0, 1], Indicates that β min represents the first preset value, β max represents the second preset value, K represents the preset number of iterations, Δm = rand(1:3f)×|step|, rand(1:3f) represents a random number between [0, 1] in the 3f dimension, and f represents the number of sampling points in each seismic trace. r1, r2, r3, r4, v are different from each other, r1, r2, r3, r4 represent four integers randomly selected from [1, N], N represents the total number of individuals in the population, m best represents the global optimal individual position, m worst represents the global worst individual position, ε represents a very small number greater than zero, ρ2=2×rand×α-α, r a and r b Represents two random numbers between [0, 1].

[0124] Optionally, the position of the target individual in the next generation is expressed as:

[0125]

[0126] Among them, f1 represents a random number in [-1,1], rand and μ1 represent random numbers between [0,1]. μ2 represents a random number between [0,1], m rand =M min +rand×(M max -M min ) indicates that Mmax represents the maximum boundary of the solution space, M min represents the minimum boundary of the solution space, represents a randomly selected individual in the current population, and f2 represents a random number from a normal distribution with a variance of 0 and a standard deviation of 1.

[0127] Example 3

[0128] See also Figure 4 An embodiment of the present invention provides an electronic device, which includes: a processor 60, a memory 61, a bus 62 and a communication interface 63, wherein the processor 60, the communication interface 63 and the memory 61 are connected via the bus 62; the processor 60 is used to execute an executable module stored in the memory 61, such as a computer program.

[0129] The memory 61 may include high-speed random access memory (RAM) and may also include non-volatile memory, such as at least one disk storage. The system network element communicates with at least one other network element via at least one communication interface 63 (which may be wired or wireless), and may utilize the Internet, a wide area network, a local area network, a metropolitan area network, or the like.

[0130] The bus 62 may be an ISA bus, a PCI bus, or an EISA bus. The bus may be divided into an address bus, a data bus, a control bus, and the like. For ease of representation, Figure 4 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0131] Among them, the memory 61 is used to store programs, and the processor 60 executes the program after receiving the execution instruction. The method executed by the device defined by the process disclosed in any embodiment of the above-mentioned embodiment of the present invention can be applied to the processor 60 or implemented by the processor 60.

[0132] The processor 60 may be an integrated circuit chip with signal processing capabilities. During implementation, the steps of the above method may be performed by hardware integrated logic circuits or software instructions within the processor 60. The processor 60 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It may implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of the present invention. The general-purpose processor may be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of the present invention may be directly executed by a hardware decoding processor or by a combination of hardware and software modules within the decoding processor. The software modules may be located in storage media well-known in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or the like. The storage medium is located in the memory 61 , and the processor 60 reads the information in the memory 61 and completes the steps of the above method in combination with its hardware.

[0133] An embodiment of the present invention provides a computer program product of a pre-stack inversion method and apparatus based on a gradient optimizer, comprising a computer-readable storage medium storing a non-volatile program code executable by a processor. The instructions included in the program code can be used to execute the method described in the previous method embodiment. For specific implementation, please refer to the method embodiment and will not be repeated here.

[0134] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0135] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present invention, 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. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0136] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings.

[0137] In the description of the present invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer" and the like indicate positions or locations based on the positions shown in the accompanying drawings, or the positions or locations in which the inventive product is typically placed when in use. These terms are intended solely to facilitate the description of the present invention and to simplify the description, and are not intended to indicate or imply that the devices or components referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limitations on the present invention. Furthermore, the terms "first," "second," and "third," etc., are used solely to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0138] Furthermore, terms such as "horizontal," "vertical," and "overhanging" do not necessarily imply that a component must be absolutely horizontal or overhanging, but rather that it can be slightly tilted. For example, "horizontal" simply means that its direction is more horizontal than "vertical," and does not mean that the structure must be completely horizontal, but rather that it can be slightly tilted.

[0139] In the description of the present invention, it should also be noted that, unless otherwise expressly specified or limited, the terms "disposed," "installed," "connected," and "connected" should be understood in a broad sense. For example, they may refer to fixed connections, detachable connections, or integral connections; they may refer to mechanical connections or electrical connections; they may refer to direct connections or indirect connections through an intermediate medium; and they may refer to internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on the specific circumstances.

[0140] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A prestack inversion method based on a gradient optimizer, characterized in that: include: Acquire pre-stack angle gather data, horizon data, and elastic parameters of a target well log in a target work area; wherein the pre-stack angle gather data is a collection of angle gather data of multiple seismic traces; and the target well log represents any well log in the target work area; Determining a solution space of elastic parameters in the target work area based on the elastic parameters of the target well logging and the horizon data of the target work area; constructing a first objective function for inverting elastic parameters in the target work area based on prestack angle gather data of the target work area and a Zoeppritz equation; Under the constraints of the solution space, a first population for solving the elastic parameters of a target seismic trace is randomly initialized; wherein the target seismic trace represents any seismic trace among the multiple seismic traces; in the first population, the position of each individual corresponds one-to-one to a potential solution of the elastic parameters of the target seismic trace; Iteratively updating the first population using a gradient optimizer to determine an optimal solution for the elastic parameters of the target seismic trace based on the angle gather data of the target seismic trace and the first objective function; constructing an elastic parameter profile of the target work area based on the optimal solutions of the elastic parameters of the multiple seismic traces; Wherein, based on the pre-stack angle gather data of the target work area and the Zoeppritz equation, a first objective function for inverting the elastic parameters in the target work area is constructed, including: Stacking the pre-stack angle gather data of the target work area according to a specified number of incident angles to obtain post-stack seismic data; extracting seismic wavelet data of the target work area from the post-stack seismic data; constructing a convolution function of elastic parameters to synthetic seismic data based on the seismic wavelet data and the Zoeppritz equation; Constructing the first objective function based on the convolution function, the target regularization parameter, the preset selection matrix and the preset scaling matrix; The first objective function is expressed as: Among them, d obs Represents the pre-stack angle gather data of the area to be inverted, d cal represents the convolution function, λ represents the target regularization parameter, f represents the number of sampling points in each seismic trace, m represents the elastic parameter to be solved, Φ j =(D j ) T ψ -1 (D j ), D j represents the preset selection matrix, with a dimension of 3*3f, and ψ represents the preset scale matrix, with a dimension of 3*3.

2. The prestack inversion method based on the gradient optimizer according to claim 1, characterized in that: The method further comprises: Constructing a second objective function for inverting elastic parameters of the target well logging based on the pre-stack angle gather data of the target well logging, the convolution function, the initial regularization parameter, the preset selection matrix and the preset scaling matrix; Under the constraints of the solution space, randomly initializing a second population for solving the elastic parameters of the target well log; wherein the position of each individual in the second population corresponds one-to-one to a potential solution of the elastic parameters of the target well log; Iteratively updating the second population using a gradient optimizer to determine an initial solution of elastic parameters of the target seismic trace based on the second objective function; When it is determined that the error between the elastic parameter of the target well logging and the initial solution does not meet the preset error requirement, the initial regularization parameter is adjusted, and the previous step is returned to be executed until the error between the initial solution and the elastic parameter of the target well logging meets the preset error requirement, thereby obtaining the target regularization parameter.

3. The prestack inversion method based on gradient optimizer according to claim 1, characterized in that: Iteratively updating the first population using a gradient optimizer to determine an optimal solution for elastic parameters of the target seismic trace based on the angle gather data of the target seismic trace and the first objective function, including: Determining a third objective function for inverting elastic parameters of the target seismic trace based on the angle gather data of the target seismic trace and the first objective function; Repeat the following steps until a preset number of iterations is reached, and use the globally optimal individual position that reaches the preset number of iterations as the elastic parameter of the target seismic trace: Determine the current global optimal individual position and the global worst individual position based on the current population and the third objective function; wherein, in the first iteration, the current population is the first population; The current position of the target individual, the current position of the random individual, the global optimal individual position, and the global worst individual position are processed based on the gradient search rule in the gradient optimizer to obtain a first displacement vector direction, a second displacement vector direction, and a third displacement vector direction of the target individual; wherein the target individual represents any individual in the current population; Determining a planned updated position of the target individual based on the first displacement vector direction, the second displacement vector direction, and the third displacement vector direction of the target individual; The planned update position, the global optimal individual position, the first displacement vector direction, the second displacement vector direction and the current position of the random individual are processed based on the local escape operator in the gradient optimizer to obtain the position of the target individual in the next generation.

4. The prestack inversion method based on gradient optimizer according to claim 3, characterized in that: The first displacement vector direction is expressed as: The second displacement vector direction is expressed as: The third displacement vector direction is expressed as: The planned updated position of the target individual is expressed as: in, Indicates the position of the vth individual in the kth iteration, randn represents a random number in the range [0, 1] that conforms to the normal distribution, ρ1 = 2 × rand × α - α, rand represents a random value in [0, 1], Indicates that β min represents the first preset value, β max represents the second preset value, K represents the preset number of iterations, Δm=rand(1:3f)×|step|, rand(1:3f) represents a random number between [0, 1] in the 3f dimension, f represents the number of sampling points in each seismic trace, r1, r2, r3, r4, v are different from each other, r1, r2, r3, r4 represent four integers randomly selected from [1, N], N represents the total number of individuals in the population, m best represents the global optimal individual position, m worst represents the global worst individual position, ε represents a very small number greater than zero, ρ2=2×rand×α-α, r a and r b Represents two random numbers between [0, 1].

5. The prestack inversion method based on gradient optimizer according to claim 4, characterized in that: The position of the target individual in the next generation is expressed as: Among them, f1 represents a random number in [-1,1], rand and μ1 represent random numbers between [0,1]. μ2 represents a random number between [0,1], m rand =M min +rand×(M max -M min ) indicates that M max represents the maximum boundary of the solution space, M min represents the minimum boundary of the solution space, m p represents an individual randomly selected from the current population, and f2 represents a random number from a normal distribution with a variance of 0 and a standard deviation of 1.

6. A prestack inversion device based on a gradient optimizer, characterized in that: An acquisition module is used to acquire pre-stack angle gather data, horizon data, and elastic parameters of a target well log in a target work area; wherein the pre-stack angle gather data is a collection of angle gather data of multiple seismic traces; and the target well log represents any well log in the target work area; a determination module, configured to determine a solution space of elastic parameters in the target work area based on the elastic parameters of the target well logging and the horizon data of the target work area; A first construction module is used to construct a first objective function for inverting elastic parameters in the target work area based on the pre-stack angle gather data of the target work area and the Zoeppritz equation; an initialization module, configured to randomly initialize a first population for solving elastic parameters of a target seismic trace under the constraints of the solution space; wherein the target seismic trace represents any seismic trace among the plurality of seismic traces; and wherein the position of each individual in the first population corresponds one-to-one to a potential solution for the elastic parameters of the target seismic trace; an updating and determining module, configured to iteratively update the first population using a gradient optimizer, so as to determine an optimal solution for the elastic parameters of the target seismic trace based on the angle gather data of the target seismic trace and the first objective function; A second construction module is configured to construct an elastic parameter profile of the target work area based on the optimal elastic parameter solutions of the multiple seismic traces; The first building block is specifically used for: Stacking the pre-stack angle gather data of the target work area according to a specified number of incident angles to obtain post-stack seismic data; extracting seismic wavelet data of the target work area from the post-stack seismic data; constructing a convolution function of elastic parameters to synthetic seismic data based on the seismic wavelet data and the Zoeppritz equation; Constructing the first objective function based on the convolution function, the target regularization parameter, the preset selection matrix and the preset scaling matrix; The first objective function is expressed as: Among them, d obs Represents the pre-stack angle gather data of the area to be inverted, d cal represents the convolution function, λ represents the target regularization parameter, f represents the number of sampling points in each seismic trace, m represents the elastic parameter to be solved, Φ j =(D j ) T ψ -1 (D j ), D j represents the preset selection matrix, with a dimension of 3*3f, and ψ represents the preset scale matrix, with a dimension of 3*3.

7. An electronic device comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, wherein: When the processor executes the computer program, the steps of the prestack inversion method based on a gradient optimizer according to any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a processor, the prestack inversion method based on a gradient optimizer according to any one of claims 1 to 5 is implemented.

Images