A DC resistivity inversion method, apparatus, equipment, and medium

Abstract

This invention provides a DC resistivity inversion method, apparatus, equipment, and medium, relating to the field of geophysical exploration technology. The method involves: acquiring observed apparent resistivity data of a target area; determining the preliminary resistivity structure corresponding to the observed apparent resistivity data; constructing a training sample dataset for the target area based on the preliminary resistivity structure; optimizing a wide learning network based on the training sample dataset and a weighted objective function to obtain a model for inversion; and using the inversion model to invert the actual resistivity data of the target area based on the observed apparent resistivity data, thereby improving the accuracy of the resistivity inversion results and achieving high-precision inversion of the underground resistivity structure.

Description

Technical Field

[0001] This invention relates to the field of geophysical exploration technology, and more specifically, to a DC resistivity inversion method, apparatus, equipment, and medium. Background Technology

[0002] Various inversion algorithms have been applied to DC resistivity inversion research. Among them, machine learning-based inversion methods can automatically extract features from training samples and establish a nonlinear mapping relationship between observed data and subsurface resistivity models, providing a new approach to improve the accuracy of DC resistivity inversion. However, the prediction accuracy of this method largely depends on the quality of the training samples. In the absence of prior geological information, it is difficult to construct a training sample set that closely approximates the actual subsurface structure. Furthermore, constructing a large number of training samples is often time-consuming and labor-intensive, and the prediction performance will significantly decrease when the sample distribution does not match the resistivity model to be inverted, thus limiting the practical application of machine learning methods in DC resistivity inversion. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide a DC resistivity inversion method, apparatus, device and medium to improve the accuracy and reliability of resistivity inversion results.

[0004] In a first aspect, this application provides a DC resistivity inversion method, including:

[0005] Acquire observed apparent resistivity data for the target region;

[0006] Based on the observed apparent resistivity data, the preliminary resistivity structure corresponding to the observed apparent resistivity data is determined; whereby the preliminary resistivity structure is the initial structural model of the true resistivity distribution in the target area.

[0007] Based on the preliminary resistivity structure, a training sample dataset for the target region is constructed.

[0008] Based on the training sample dataset and the weighted objective function, the wide learning network is optimized to obtain the inversion model. The weighted objective function is determined based on the relationship between the input matrix and output weight matrix of the wide learning network, the weight coefficients of each training sample in the training sample dataset, and the correlation coefficients between the observed data and the observed apparent resistivity data in the training sample dataset.

[0009] Based on observed apparent resistivity data, an inversion model is used to retrieve the true resistivity data of the target area.

[0010] Optionally, based on the observed apparent resistivity data, a preliminary resistivity structure corresponding to the observed apparent resistivity data is determined, including:

[0011] Based on observed apparent resistivity data, iterative operations are performed on the resistivity model until the iteration termination condition is met. Then, based on the parameters of the resistivity model updated during the last iteration, a preliminary resistivity structure is obtained. The resistivity model is a mathematical model of the spatial distribution characteristics of subsurface resistivity in the target area. The iterative operations include:

[0012] Forward modeling was performed on the resistivity model to obtain theoretical apparent resistivity data;

[0013] Based on theoretical and observed apparent resistivity data, the objective function value is obtained; based on the objective function value, the gradient of the objective function value is determined; based on the gradient, the search direction of the objective function value is determined; based on the linear search criterion, the step size of the search direction is determined.

[0014] The parameters of the resistivity model are updated based on the search direction and step size.

[0015] Optionally, the iteration termination condition is that the number of iterations is not less than a threshold number, or the fitting difference is not higher than a preset error threshold, where the fitting difference is the weighted sum of squared residuals between the observed apparent resistivity data and the theoretical apparent resistivity data.

[0016] Optionally, based on the preliminary resistivity structure, a training sample dataset for the target region is constructed, including:

[0017] Statistical analysis is performed based on the preliminary resistivity structure to determine the background resistivity of the preliminary resistivity structure; whereby the background resistivity is the resistivity value that appears most frequently in the preliminary resistivity structure.

[0018] Based on the preliminary resistivity structure and background resistivity, anomalous regions of the preliminary resistivity structure are determined using a uniform random distribution function and staged clustering. The anomalous regions are determined based on the relationship between the preliminary resistivity structure and the background resistivity. Based on the background resistivity and the anomalous regions, a training sample dataset of the target region is constructed.

[0019] Optionally, a training sample dataset for the target region is constructed based on background resistivity and anomalous body regions, including:

[0020] The initial training dataset is obtained by processing the background resistivity and anomalous body regions using random distribution.

[0021] Based on the initial training dataset, the initial observation dataset is obtained by forward modeling.

[0022] The initial observation dataset is standardized and scaled to obtain a standardized observation dataset.

[0023] Based on the observed apparent resistivity data and the initial training dataset, the target training dataset is obtained;

[0024] A training sample dataset for the target region is constructed based on the target training dataset and the standardized observation dataset.

[0025] Optionally, the DC resistivity inversion method provided in this application further includes:

[0026] The correlation coefficient was calculated based on observed apparent resistivity data using covariance and standard deviation.

[0027] The prediction error is obtained by comparing the input matrix and output weight matrix of the width learning network with the training dataset in the training sample dataset.

[0028] Based on the input matrix and output weight matrix of the width learning network, the regularization norm term is determined;

[0029] Based on the correlation coefficient, prediction error, and regularization norm term, the weighted objective function is obtained.

[0030] Optionally, based on observed apparent resistivity data, an inversion model is used to retrieve the true resistivity data of the target region, including:

[0031] The observed apparent resistivity data are standardized and scaled to obtain optimized observed apparent resistivity data.

[0032] Based on optimized observed apparent resistivity data, an inversion model is used to retrieve the true resistivity data of the target region.

[0033] Secondly, this application provides a DC resistivity inversion device, comprising:

[0034] The data acquisition module is used to acquire observed apparent resistivity data of the target area;

[0035] The data processing module is used to determine the preliminary resistivity structure corresponding to the observed apparent resistivity data based on the observed apparent resistivity data; and to construct a training sample dataset for the target region based on the preliminary resistivity structure; wherein, the preliminary resistivity structure is the initial structural model of the true resistivity distribution of the target region.

[0036] The model training module is used to optimize the wide learning network based on the training sample dataset and the weighted objective function to obtain the inversion model. The weighted objective function is determined based on the relationship between the input matrix and output weight matrix of the wide learning network, the weight coefficients of each training sample in the training sample dataset, and the correlation coefficients between the observed data and the observed apparent resistivity data in the training sample dataset.

[0037] The data inversion module is used to invert the true resistivity data of the target area based on the observed apparent resistivity data and using an inversion model.

[0038] Thirdly, this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described DC resistivity inversion method.

[0039] Fourthly, this application provides a computer-readable storage medium storing computer instructions that, when executed by a processor, implement the aforementioned DC resistivity inversion method.

[0040] This invention provides a DC resistivity inversion method, apparatus, device, and medium. The method involves: acquiring observed apparent resistivity data of a target region; determining the preliminary resistivity structure corresponding to the observed apparent resistivity data; constructing a training sample dataset for the target region based on the preliminary resistivity structure; optimizing a wide learning network based on the training sample dataset and a weighted objective function to obtain an inversion model; and using the inversion model to invert the actual resistivity data of the target region based on the observed apparent resistivity data, thereby improving the accuracy of the resistivity inversion results and achieving high-precision inversion of the underground resistivity structure.

[0041] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0042] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 A flowchart of a DC resistivity inversion method provided by an embodiment of the present invention is shown;

[0044] Figure 2 A schematic diagram of the structure of the synthetic theoretical resistivity model provided in an embodiment of the present invention is shown;

[0045] Figure 3 A schematic diagram of the preliminary resistivity structure provided in an embodiment of the present invention is shown;

[0046] Figure 4 This diagram illustrates the process of obtaining the initial resistivity structure provided by an embodiment of the present invention.

[0047] Figure 5 A statistical histogram of the preliminary resistivity structure provided in an embodiment of the present invention is shown;

[0048] Figure 6 A schematic diagram of multiple abnormal body regions provided in an embodiment of the present invention is shown;

[0049] Figure 7 A schematic diagram of the training sample dataset provided in an embodiment of the present invention is shown;

[0050] Figure 8 A schematic diagram of the actual resistivity data provided in the embodiments of the present invention is shown;

[0051] Figure 9 A schematic diagram of the structure of a DC resistivity inversion device provided in an embodiment of the present invention is shown;

[0052] Figure 10 A schematic diagram of the structure of an electronic device provided in an embodiment of the present invention is shown. Detailed Implementation

[0053] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. 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 claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0054] This application provides a method for inverting DC resistivity, see below. Figure 1 As shown in the embodiments of this application, the general flow of the DC resistivity inversion method is as follows:

[0055] Step 110: Obtain the observed apparent resistivity data of the target area.

[0056] In this embodiment of the application, the observed apparent resistivity data of the target area can be obtained in the following ways:

[0057] The observed apparent resistivity data of the target region is obtained by performing forward modeling on the resistivity model of the target region; wherein, the resistivity model of the target region can be a synthetic theoretical resistivity model or an actual geological model. Further, in this embodiment, the observed apparent resistivity data is obtained by performing forward modeling on a synthetic theoretical resistivity model, such as... Figure 2 The model shown is a composite theoretical resistivity model, with each black dot representing an electrode for each test line.

[0058] Step 120: Based on the observed apparent resistivity data, determine the preliminary resistivity structure corresponding to the observed apparent resistivity data; wherein, the preliminary resistivity structure is the initial structural model of the true resistivity distribution in the target area.

[0059] In this embodiment of the application, the first-stage inversion of the observed apparent resistivity data is performed using a finite-memory quasi-Newton method to obtain the following: Figure 3 The preliminary resistivity structure shown, wherein, as Figure 4 The specific process for obtaining the initial resistivity structure is as follows:

[0060] Step 121: Perform forward modeling on the resistivity model to obtain theoretical apparent resistivity data.

[0061] Step 122: Based on theoretical and observed apparent resistivity data, obtain the objective function value; based on the objective function value, determine the gradient of the objective function value; based on the gradient, determine the search direction of the objective function value.

[0062] In this embodiment, a weighted data residual term is obtained based on observed apparent resistivity data; a regularization term is obtained based on the resistivity model and a reference model (resistivity model or uniform background model); and an objective function value is obtained based on the data residual term and the regularization term.

[0063] Based on the objective function value, the Jacobian matrix is ​​obtained by taking the partial derivative of the objective function value through the perturbation method or the adjoint method; the gradient of the data residual term is obtained based on the Jacobian matrix and the data residual term; the gradient of the regularization term is obtained based on the Jacobian matrix and the regularization term; the gradient of the objective function value is determined based on the gradient of the data residual term and the gradient of the regularization term.

[0064] By combining the model changes and gradient changes from historical iterations with quasi-Newton conditions, the gradient is updated using a two-loop recursive algorithm to obtain the search direction.

[0065] Step 123: Determine the step size of the search direction based on the linear search criterion.

[0066] In this embodiment of the application, the step size for determining the search direction can be determined by the Armijo (Armijo Condition) criterion or the Wolfe (Wolfe Conditions) criterion.

[0067] Step 124: Update the parameters of the initial resistivity based on the search direction and step size.

[0068] Step 125: Determine if the iteration termination condition is met; if yes, proceed to step 126; if no, return to step 121; wherein, the iteration termination condition is that the number of iterations is not less than the number threshold, or the fitting difference is not higher than the preset error threshold, and the fitting difference is the weighted sum of squared residuals of the observed apparent resistivity data and the theoretical apparent resistivity data.

[0069] Step 126: Based on the parameters of the resistivity model updated during the last iteration, obtain the preliminary resistivity structure.

[0070] Step 130: Based on the preliminary resistivity structure, construct the training sample dataset for the target region.

[0071] In this embodiment of the application, the training sample dataset for the target region can be determined in the following ways:

[0072] Statistical analysis is performed based on the preliminary resistivity structure to determine the background resistivity of the preliminary resistivity structure. The background resistivity is the resistivity value that appears most frequently in the preliminary resistivity structure. Based on the preliminary resistivity structure and the background resistivity, anomalous regions of the preliminary resistivity structure are determined using a uniform random distribution function and staged clustering. The anomalous regions are determined based on the relationship between the preliminary resistivity structure and the background resistivity.

[0073] Furthermore, based on the background resistivity and anomalous body regions, a training sample dataset for the target region is constructed, including: obtaining an initial training dataset by processing the background resistivity and anomalous body regions using a random distribution; obtaining an initial observation dataset based on the initial training dataset; standardizing and scaling the initial observation dataset to obtain a standardized observation dataset; obtaining a target training dataset based on the observed apparent resistivity data and the initial training dataset; and constructing a training sample dataset for the target region based on the target training dataset and the standardized observation dataset, wherein the anomalous body regions include high-resistivity anomalous body regions and low-resistivity anomalous body regions, and the resistivity is initially structured. The region is defined as a low-resistivity anomaly region; the preliminary resistivity structure is then defined. The region is defined as a high-resistivity anomaly region.

[0074] In practical implementation, the specific process of determining the training sample dataset for the target region based on the preliminary resistivity structure is as follows:

[0075] First, a statistical analysis was performed on the preliminary resistivity structure, and based on the results of the statistical analysis, the following was obtained: Figure 5The statistical histogram of the preliminary resistivity structure is shown, where the horizontal axis represents the resistivity value and the vertical axis represents the number of times each resistivity value appears in the preliminary resistivity structure. The resistivity value with the highest frequency of occurrence in the statistical histogram of the preliminary resistivity structure is determined as the background resistivity of the preliminary resistivity structure.

[0076] Secondly, based on the resistivity of each element in the preliminary resistivity structure and the background resistivity, determine the maximum value of the relative change between the preliminary resistivity structure and the background resistivity:

[0077]

[0078] In the formula, This represents the maximum value of the relative change between the initial resistivity structure and the background resistivity. This is the preliminary resistivity structure. Background resistivity;

[0079] like Figure 5 The statistical histogram of the preliminary resistivity structure shows that the number of grids with a resistivity value of 100 ohm-meters is the largest. Therefore, the background resistivity can be determined to be 100 ohm-meters. The maximum value RE of the relative change between the preliminary resistivity structure and the background resistivity can be calculated using the above formula as 4.77769.

[0080] Based on the maximum relative change between the preliminary resistivity structure and the background resistivity, N thresholds are generated using a uniform random distribution function. In the formula, delta is the threshold, and rand(N) is a random number between 0 and 1. Based on the threshold, a staged clustering strategy is used to identify and cluster the preliminary resistivity structure, resulting in low-resistivity anomaly regions and high-resistivity anomaly regions. To avoid misclassifying low-resistivity and high-resistivity anomaly regions as the same class during spatial clustering, this embodiment uses the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm to cluster the low-resistivity regions in the preliminary resistivity structure, resulting in multiple low-resistivity anomaly regions. The DBSCAN algorithm is also used to cluster the high-resistivity regions in the preliminary resistivity structure, resulting in multiple high-resistivity anomaly regions, such as... Figure 6 As shown, the average resistivity of the low-resistivity anomaly region was determined. and the average resistivity of the high-resistivity anomaly region ;

[0081] The training sample dataset is obtained based on the background resistivity, various low-resistivity anomaly regions, and various high-resistivity anomaly regions. First, an initial training dataset is obtained based on the background resistivity, various low-resistivity anomaly regions, and various high-resistivity anomaly regions. The resistivity value of the low-resistivity anomaly in the low-resistivity anomaly regions of the initial training dataset is... Random numbers uniformly distributed between; the resistivity value of the high-resistivity anomaly in the high-resistivity anomaly region is Random numbers between;

[0082] Secondly, forward modeling is performed on the training data in the initial training dataset to obtain the initial observation dataset; then, the initial observation dataset is standardized and scaled using the following formula to obtain the standardized observation dataset, thereby improving the sensitivity to the resistivity values ​​of high-resistivity anomalies and low-resistivity anomalies:

[0083]

[0084] In the formula, To standardize the observation dataset, For the initial observation dataset, The mean, Standard deviation;

[0085] Then, the initial training dataset is transformed to obtain the transformed training dataset, i.e. This is done to alleviate the problem of excessively large ranges in resistivity data; and the observed apparent resistivity data are preprocessed using the following formula:

[0086]

[0087] In the formula, The observed apparent resistivity data after preprocessing. To observe apparent resistivity data, The mean, Standard deviation;

[0088] The correlation coefficient between the preprocessed observed apparent resistivity data and the training data in the transformed training dataset is determined using the following formula, and the training data with high similarity in the transformed training dataset are selected as the target training dataset:

[0089]

[0090] In the formula, The correlation coefficient is... For covariance, Standard deviation;

[0091] Finally, the training sample dataset is obtained based on the target training dataset and the normalized observation dataset, such as... Figure 7 As shown.

[0092] Step 140: Optimize the wide learning network based on the training sample dataset and the weighted objective function to obtain the inversion model; wherein, the weighted objective function is determined based on the relationship between the input matrix and output weight matrix of the wide learning network, the weight coefficients of each training sample in the training sample dataset, and the correlation coefficients between the observed data and the observed apparent resistivity data in the training sample dataset.

[0093] In this embodiment, the weighted objective function can be obtained in the following way:

[0094] Based on observed apparent resistivity data, the correlation coefficient is calculated using covariance and standard deviation; the prediction error is obtained based on the difference between the input matrix and output weight matrix of the wide-range learning network and the training dataset; the regularization norm term is determined based on the input matrix and output weight matrix of the wide-range learning network; and the weighted objective function is obtained based on the correlation coefficient, prediction error, and regularization norm term, where the weighted objective function is:

[0095]

[0096] In the formula, For exponential operations, The correlation coefficient is... Here, A represents the weight coefficients for each training sample, A is the input matrix of the width learning network (i.e., the target training dataset), and W is the weight matrix of the width learning network. where is the weight coefficient of the regularization term; By ensuring that samples with higher correlation coefficients receive greater weight values, they can significantly influence the final results during the training process.

[0097] Step 150: Based on the observed apparent resistivity data, use the inversion model to invert the true resistivity data of the target area.

[0098] In this embodiment, the observed apparent resistivity data is standardized and scaled to obtain optimized observed apparent resistivity data; based on the optimized observed apparent resistivity data, an inversion model is used to retrieve the true resistivity data of the target region, as follows: Figure 8 As shown; among them, the observed apparent resistivity data is standardized and scaled to obtain optimized observed apparent resistivity data, using the following formula:

[0099]

[0100] Furthermore, to quantitatively evaluate the accuracy of the results obtained by the DC resistivity inversion method provided in this application, the structural similarity coefficient (SSIM) is also used in the embodiments of this application to compare the differences between the inversion results and the theoretical resistivity model. The SSIM of the preliminary resistivity structure result and the theoretical model is 0.3774, while the SSIM of the inversion model result and the theoretical resistivity model is 0.7988. The comparison results show that the DC resistivity inversion method proposed in this invention can significantly improve the accuracy of the inversion results.

[0101] The DC resistivity inversion method provided in this application integrates the finite-memory quasi-Newton method with width learning for DC resistivity inversion, which ensures the reliability of the inversion results and improves the efficiency of the inversion. The staged DBSCAN algorithm can automatically identify low-resistivity and high-resistivity anomalous regions, avoiding manual division and improving the intelligence of sample generation. The two-stage inversion can effectively avoid the problem of traditional optimization getting trapped in local extrema, while reducing the dependence of machine learning methods on a large number of samples.

[0102] This application provides a DC resistivity inversion device, see [link / reference]. Figure 9 As shown, the DC resistivity inversion device provided in this application includes:

[0103] Data acquisition module 910 is used to acquire observed apparent resistivity data of the target area;

[0104] The data processing module 920 is used to determine the preliminary resistivity structure corresponding to the observed apparent resistivity data based on the observed apparent resistivity data; and to construct a training sample dataset for the target region based on the preliminary resistivity structure; wherein, the preliminary resistivity structure is the initial structural model of the true resistivity distribution of the target region.

[0105] The model training module 930 is used to optimize the wide learning network based on the training sample dataset and the weighted objective function to obtain the inversion model. The weighted objective function is determined based on the relationship between the input matrix and output weight matrix of the wide learning network, the weight coefficients of each training sample in the training sample dataset, and the correlation coefficients between the observed data and the observed apparent resistivity data in the training sample dataset. The data inversion module 940 is used to invert the true resistivity data of the target region based on the observed apparent resistivity data and the inversion model.

[0106] It should be noted that the principle of the DC resistivity inversion device provided in this application embodiment to solve the technical problem is similar to that of the DC resistivity inversion method provided in this application embodiment. Therefore, the implementation of the DC resistivity inversion device provided in this application embodiment can refer to the implementation of the DC resistivity inversion method provided in this application embodiment, and the repeated parts will not be described again.

[0107] After introducing the DC resistivity inversion method and apparatus provided in the embodiments of this application, the electronic equipment provided in the embodiments of this application will be briefly introduced next.

[0108] See Figure 10 As shown, the electronic device 500 provided in this application embodiment includes at least a processor 501, a memory 502, and a computer program stored in the memory 502 and executable on the processor 501. When the processor 501 executes the computer program, it implements the DC resistivity inversion method provided in this application embodiment.

[0109] The electronic device 500 provided in this application embodiment may further include a bus 503 connecting different components (including processor 501 and memory 502). The bus 503 represents one or more types of bus structures, including memory bus, peripheral bus, local area bus, etc.

[0110] Memory 502 may include a readable storage medium in the form of volatile memory, such as random access memory (RAM) 5021 and / or cache memory 5022, and may further include read-only memory (ROM) 5023. Memory 502 may also include a program tool 5025 having a set (at least one) of program modules 5024, including but not limited to an operating subsystem, one or more application programs, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment.

[0111] Processor 501 can be a single processing element or a collective term for multiple processing elements. For example, processor 501 can be a central processing unit (CPU) or one or more integrated circuits configured to implement the DC resistivity inversion method provided in the embodiments of this application. Specifically, processor 501 can be a general-purpose processor, including but not limited to CPUs, application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0112] Electronic device 500 can communicate with one or more external devices 504 (e.g., keyboard, remote control, etc.), and also with one or more devices that enable a user to interact with electronic device 500 (e.g., mobile phone, computer, etc.), and / or with devices that enable electronic device 500 to communicate with one or more other electronic devices 500 (e.g., router, modem, etc.). This communication can be performed through input / output (I / O) interface 505. Furthermore, electronic device 500 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) through network adapter 506. Figure 10 As shown, network adapter 506 communicates with other modules of electronic device 500 via bus 503. It should be understood that, although... Figure 10 As not shown, other hardware and / or software modules may be used in conjunction with the electronic device 500, including but not limited to microcode, device drivers, redundant processors, external disk drive arrays, Redundant Arrays of Independent Disks (RAID) subsystems, tape drives, and data backup storage subsystems.

[0113] It should be noted that, Figure 10 The electronic device 500 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0114] The computer-readable storage medium provided in the embodiments of this application is described below. The computer-readable storage medium provided in the embodiments of this application stores computer instructions, which, when executed by a processor, implement the DC resistivity inversion method provided in the embodiments of this application. Specifically, the computer instructions can be built into or installed in the processor, so that the processor can implement the DC resistivity inversion method provided in the embodiments of this application by executing the built-in or installed computer instructions.

[0115] In addition, the DC resistivity inversion method provided in this application embodiment can also be implemented as a computer program product, which includes program code. The program code implements the DC resistivity inversion method provided in this application embodiment when it is run on a processor.

[0116] The computer program product provided in this application embodiment may employ one or more computer-readable storage media, which may be, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination thereof. Specifically, more specific examples (a non-exhaustive list) of computer-readable storage media include electrical connections with one or more wires, portable disks, hard disks, RAM, ROM, erasable programmable read-only memory (EPROM), optical fibers, portable compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0117] The computer program product provided in this application embodiment can be a CD-ROM and include program code, and can also run on electronic devices such as computers. However, the computer program product provided in this application embodiment is not limited thereto. In this application embodiment, the computer-readable storage medium can be any tangible medium that contains or stores program code, which can be used by or in conjunction with an instruction execution system, device, or apparatus.

[0118] It should be noted that although several units or sub-units of the device have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of this application, the features and functions of two or more units described above can be embodied in one unit. Conversely, the features and functions of one unit described above can be further divided and embodied by multiple units.

[0119] Furthermore, although the operations of the method of this application are described in a specific order in the accompanying drawings, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.

[0120] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0121] Obviously, those skilled in the art can make various modifications and variations to the embodiments of this application without departing from the spirit and scope of the embodiments of this application. Therefore, if these modifications and variations to the embodiments of this application fall within the scope of the claims of this application and their equivalents, this application also intends to include these modifications and variations.

Claims

1. A method for inverting DC resistivity, characterized in that, include: Acquire observed apparent resistivity data for the target region; Based on the observed apparent resistivity data, a preliminary resistivity structure corresponding to the observed apparent resistivity data is determined; wherein, the preliminary resistivity structure is an initial structural model of the true resistivity distribution of the target area; determining the preliminary resistivity structure based on the observed apparent resistivity data includes: performing an iterative operation on the resistivity model based on the observed apparent resistivity data until the iteration termination condition is met, and obtaining the preliminary resistivity structure based on the parameters of the resistivity model updated during the last iteration operation; wherein, the resistivity model is a mathematical model of the spatial distribution characteristics of underground resistivity in the target area; the iterative operation includes: performing forward modeling on the resistivity model to obtain theoretical apparent resistivity data; obtaining an objective function value based on the theoretical apparent resistivity data and the observed apparent resistivity data; determining the gradient of the objective function value based on the objective function value; determining the search direction of the objective function value based on the gradient; determining the step size of the search direction based on a linear search criterion; and updating the parameters of the resistivity model based on the search direction and the step size; Based on the preliminary resistivity structure, a training sample dataset for the target region is constructed. This construction includes: performing statistical analysis to determine the background resistivity of the preliminary resistivity structure; wherein the background resistivity is the resistivity value that appears most frequently in the preliminary resistivity structure; determining anomalous regions of the preliminary resistivity structure using a uniform random distribution function and staged clustering based on the preliminary resistivity structure and the background resistivity; the anomalous regions are determined based on the relationship between the preliminary resistivity structure and the background resistivity; and constructing the training sample dataset for the target region based on the background resistivity and the anomalous regions. Based on the training sample dataset and the weighted objective function, the width learning network is optimized to obtain an inversion model; wherein, the weighted objective function is determined based on the relationship between the input matrix and output weight matrix of the width learning network, the weight coefficients of each training sample in the training sample dataset, and the correlation coefficients between the observed data and the observed apparent resistivity data in the training sample dataset; based on the observed apparent resistivity data, the inversion model is used to invert the true resistivity data of the target region.

2. The DC resistivity inversion method according to claim 1, characterized in that, The iteration termination condition is that the number of iterations is not less than a threshold number, or the fitting difference is not higher than a preset error threshold, wherein the fitting difference is the weighted sum of squared residuals between the observed apparent resistivity data and the theoretical apparent resistivity data.

3. The DC resistivity inversion method according to claim 1, characterized in that, Based on the background resistivity and the anomalous body region, a training sample dataset for the target region is constructed, including: Based on the background resistivity and the anomalous body region, an initial training dataset is obtained by processing with random distribution. Based on the initial training dataset, the initial observation dataset is obtained by forward modeling. The initial observation dataset is standardized and scaled to obtain a standardized observation dataset; Based on the observed apparent resistivity data and the initial training dataset, the target training dataset is obtained; Based on the target training dataset and the standardized observation dataset, a training sample dataset for the target region is constructed.

4. The DC resistivity inversion method according to claim 3, characterized in that, Also includes: Based on the observed apparent resistivity data, the correlation coefficient was calculated using covariance and standard deviation; The prediction error is obtained based on the difference between the input matrix and output weight matrix of the wide learning network and the training dataset in the training sample dataset; Based on the input matrix and output weight matrix of the wide learning network, the regularization norm term is determined; The weighted objective function is obtained based on the correlation coefficient, the prediction error, and the regularization norm term.

5. The DC resistivity inversion method according to claim 1, characterized in that, Based on the observed apparent resistivity data, the inversion model is used to invert the true resistivity data of the target region, including: The observed apparent resistivity data is standardized and scaled to obtain optimized observed apparent resistivity data; Based on the optimized observed apparent resistivity data, the inversion model is used to invert the true resistivity data of the target region.

6. A DC resistivity inversion device, characterized in that, include: The data acquisition module is used to acquire observed apparent resistivity data of the target area; The data processing module is used to determine the preliminary resistivity structure corresponding to the observed apparent resistivity data based on the observed apparent resistivity data; and to construct a training sample dataset for the target region based on the preliminary resistivity structure. The preliminary resistivity structure is an initial structural model of the true resistivity distribution of the target region. Determining the preliminary resistivity structure based on the observed apparent resistivity data includes: performing an iterative operation on the resistivity model based on the observed apparent resistivity data until the iteration termination condition is met; and obtaining the preliminary resistivity structure based on the parameters of the resistivity model updated during the last iteration. The resistivity model is a mathematical model of the spatial distribution characteristics of subsurface resistivity in the target region. The iterative operation includes: performing forward modeling on the resistivity model to obtain theoretical apparent resistivity data; and obtaining the target function based on the theoretical apparent resistivity data and the observed apparent resistivity data. The process involves: determining the gradient of the objective function value based on the objective function value; determining the search direction of the objective function value based on the gradient; determining the step size of the search direction based on a linear search criterion; updating the parameters of the resistivity model based on the search direction and the step size; and constructing a training sample dataset for the target region based on the preliminary resistivity structure, including: determining the background resistivity of the preliminary resistivity structure through statistical analysis; wherein the background resistivity is the resistivity value that appears most frequently in the preliminary resistivity structure; determining the anomalous regions of the preliminary resistivity structure based on the preliminary resistivity structure and the background resistivity using a uniform random distribution function and staged clustering; the anomalous regions are determined based on the relationship between the preliminary resistivity structure and the background resistivity; and constructing a training sample dataset for the target region based on the background resistivity and the anomalous regions. The model training module is used to optimize the wide learning network based on the training sample dataset and the weighted objective function to obtain the inversion model; wherein, the weighted objective function is determined based on the relationship between the input matrix and output weight matrix of the wide learning network, the weight coefficients of each training sample in the training sample dataset, and the correlation coefficients between the observed data and the observed apparent resistivity data in the training sample dataset; The data inversion module is used to invert the actual resistivity data of the target area based on the observed apparent resistivity data and using the inversion model.

7. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the DC resistivity inversion method as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the DC resistivity inversion method as described in any one of claims 1 to 5.

Images