A method for three-dimensional modeling and visualization based on gravity and magnetic data

Abstract

This invention discloses a three-dimensional modeling and visualization method based on gravity and magnetic data, relating to the field of geophysical exploration technology. The method includes the following steps: performing multi-scale decomposition processing on gravity and magnetic anomaly data to weaken the volume field superposition effect; constructing a three-dimensional grid model of the study area based on the spatial distribution of surface observation data, using the density and magnetic susceptibility of each grid cell as model variables to be inverted, and reducing the impact of superposition effect on subsurface exploration modeling through multiple gravity and magnetic field components with different depth sensitivities, providing high-quality input for the constructed three-dimensional grid model; then inverting the three-dimensional grid model, and introducing differentiated weight constraints during the inversion process to improve the resolution and stability of the subsurface physical property distribution inversion, and enhance the accuracy of inferring subsurface geological properties.

Description

Technical Field

[0001] This invention relates to the field of geophysical surveying technology, specifically to a three-dimensional modeling and visualization method based on gravity and magnetic data. Background Technology

[0002] Gravity and magnetic exploration is an important technical means for mineral resource exploration, regional geological surveys and deep geological structure research. It mainly uses gravity anomaly data and magnetic anomaly data as the core to study geophysical exploration work. By observing the variation patterns of gravity field and geomagnetic field, geophysical exploration methods can infer the underground geological structure and mineral distribution.

[0003] In gravity and magnetic exploration, the observed gravity and magnetic anomalies are the superposition results of responses caused by the differences in physical properties of different underground geological bodies. This is because the gravity and magnetic fields have volume field characteristics, and the contributions of geological bodies at different depths and spatial locations to the observation data are coupled with each other, resulting in obvious non-uniqueness and multiple solutions between the observation data and the distribution of underground physical properties. Therefore, it is difficult to accurately recover the distribution of underground density and magnetic susceptibility, especially when multi-scale geological bodies are superimposed or when the differences in geological physical properties are not significant. Traditional inversion methods rely on feature extraction and modeling of gravity and magnetic data, which makes it difficult to separate the contributions of anomalies at different depth levels, thus affecting the accuracy of the inversion 3D modeling. Summary of the Invention

[0004] I. Technical problems to be solved Therefore, this invention provides a three-dimensional modeling and visualization method based on gravity and magnetic data, which can solve the influence of superposition of gravity and magnetic fields on the inversion modeling of exploration data.

[0005] II. Technical Solution To achieve the above objectives, the present invention provides the following technical solution: a three-dimensional modeling and visualization method based on gravity and magnetic data, the method comprising the following steps: Surface observation data of the study area are obtained, including gravity anomaly data and magnetic anomaly data. The gravity anomaly data is the observation result of the density distribution of the subsurface medium after gravity forward modeling response, and the magnetic anomaly data is the observation result of the magnetic susceptibility distribution of the subsurface medium after magnetic field forward modeling response. The gravity anomaly data is preprocessed with zero-point drift correction and format conversion, and the magnetic anomaly data is preprocessed with denoising and polarization. The preprocessed gravity and magnetic anomaly data are subjected to multi-scale decomposition to weaken the volume field superposition effect and obtain gravity and magnetic field components reflecting different depths underground. The multi-scale decomposition includes extracting low-frequency components of deep underground regions through upward extension processing, extracting local anomaly components of shallow underground regions through high-pass filtering, and extracting anomaly boundaries that enhance the signal. Based on the spatial distribution of surface observation data, a three-dimensional grid model of the study area is constructed, the underground space is divided into multiple continuous grid units, and the density parameter and magnetic susceptibility parameter of each grid unit are used as model variables to be inverted. Initial values ​​are assigned to the model variables to construct an initial model. The gravity response and magnetic field response corresponding to the initial model are calculated by gravity forward modeling operator and magnetic field forward modeling operator, and compared with the observed gravity anomaly data and magnetic anomaly data to obtain the corresponding gravity residual and magnetic residual; A comprehensive inversion objective function for gravity and magnetic fields is constructed, which includes gravity data fitting terms, magnetic data fitting terms, and model regularization constraint terms. Weight parameters for different fitting terms are set using the gravity and magnetic field components. Based on the gravity residuals and magnetic field residuals, the comprehensive inversion objective function is iteratively solved using the conjugate gradient method to obtain the underground density distribution model and the magnetic susceptibility distribution model, which are then visualized in three dimensions.

[0006] Furthermore, surface observation data of the study area is obtained. Multiple surface test points with different geographical locations are preset in the study area. Gravity anomaly data of the surface observation data are collected at the surface test points using a ground gravimeter, and magnetic anomaly data of the surface observation data are collected at the surface test points using a magnetometer.

[0007] Furthermore, the preprocessed gravity anomaly data and magnetic anomaly data undergo multi-scale decomposition processing, including the following steps: The gravity and magnetic anomaly data are gridded to construct a regular spatial data field; Based on the gridded data, the gravity anomaly data and magnetic anomaly data are extended at different heights using the up-extension method to obtain low-frequency field components corresponding to different extension heights, which are used to characterize the structural information of deep underground layers. High-pass filtering was performed on gravity anomaly data and magnetic anomaly data to separate high-frequency anomaly components, which were used to characterize local anomaly information in shallow underground layers. Vertical derivatives and analytical signals are calculated for gravity anomaly data and magnetic anomaly data to enhance the response of anomaly boundaries and extract gradient features that reflect the boundaries of geological bodies.

[0008] Furthermore, when constructing a three-dimensional mesh model of the study area, the boundary range of the underground modeling space is first determined based on the spatial distribution range of the surface observation data, including the horizontal range and the vertical depth range; After determining the spatial range, the underground space is divided into multiple three-dimensional grid units according to the preset spatial resolution. Corresponding physical property parameters are defined for each grid unit. The density parameter and magnetic susceptibility parameter of each grid unit are set as model variables to be solved by inversion, which are used to characterize the physical property distribution characteristics of different underground locations.

[0009] Furthermore, after obtaining the initial model, based on the three-dimensional mesh model and model variables, forward modeling calculations of the gravitational and magnetic fields are performed, and the following steps are executed: For each observation point on the surface, the gravity response of all underground grid cells to that observation point is calculated, and the contributions of all grid cells are summed to obtain the predicted gravity value corresponding to that observation point. Based on the magnetic susceptibility parameters of each grid cell, the magnetic field response generated at each observation point is calculated and accumulated to obtain the predicted magnetic field value corresponding to each observation point. Obtain the predicted gravity data and predicted magnetic field data corresponding to each observation point.

[0010] Furthermore, after obtaining the predicted gravity data and predicted magnetic field data, they are compared with the observed gravity anomaly data and magnetic anomaly data on a grid-by-grid basis, and the difference at each observation point is calculated; wherein, the difference between the predicted gravity data and the observed gravity anomaly data is taken as the gravity residual, and the difference between the predicted magnetic field data and the observed magnetic anomaly data is taken as the magnetic field residual.

[0011] Furthermore, gravity and magnetic data inversion is performed based on the aforementioned three-dimensional mesh model. A comprehensive inversion objective function is constructed, integrating gravity data fitting terms, magnetic data fitting terms, and model regularization constraint terms. The magnetic data fitting term is generated through a nonlinear equation, specifically: ; in, This represents the total number of magnetic observation locations. For the magnetic anomaly data collected at the i-th observation location in the study area, The predicted magnetic field data is obtained from the forward modeling of the three-dimensional mesh model at the i-th observation location. The weighting parameters for magnetic anomaly data. It is a constant.

[0012] Furthermore, a comprehensive inversion objective function is constructed that integrates the gravity data fitting term, the magnetic data fitting term, and the model regularization constraint term. The comprehensive inversion objective function is as follows: ; in, For the gravity data fitting term, For the magnetic data fitting term, For the physical property constraints of the model.

[0013] Furthermore, the comprehensive inversion objective function performs joint inversion of the physical properties of the gravity and magnetic spatial characteristics at different scales. Different weighting parameters for fitting terms are set using the gravity and magnetic field components. Based on the gravity residual and magnetic field residual, the comprehensive inversion objective function is iteratively minimized using the conjugate gradient method. In each iteration, the gradient vector of the comprehensive inversion objective function with respect to density and magnetic susceptibility is calculated. The conjugate direction is determined using the gradient data from the previous iteration, and the search is performed along the conjugate direction to determine the iteration step size, thereby obtaining the density distribution model and magnetic susceptibility distribution model of the study area.

[0014] III. Beneficial Effects: Compared with the prior art, this invention has the following beneficial effects: This invention provides a three-dimensional modeling and visualization method based on gravity and magnetic data. The gravity and magnetic fields acquired in the study area are decomposed into multiple components with different depth sensitivities. Differentiated weight constraints are introduced during the inversion process to effectively separate the low-frequency interference background field in deep underground space and the local anomalies in shallow underground space. This reduces the impact of superposition effects on underground exploration modeling and improves the resolution and stability of underground property distribution inversion. Attached Figure Description

[0015] Figure 1 This is a working block diagram of a three-dimensional modeling and visualization method based on gravity and magnetic data according to the present invention; Figure 2 A flowchart illustrating a three-dimensional modeling and visualization method based on gravity and magnetic data provided in an embodiment of the present invention; Figure 3 A schematic diagram of a density distribution model for a three-dimensional modeling and visualization method based on gravity and magnetic data provided in an embodiment of the present invention; Figure 4 A schematic diagram of a magnetic susceptibility distribution model based on a three-dimensional modeling and visualization method using gravity and magnetic data, provided in an embodiment of the present invention; Figure 5 A schematic diagram of a three-dimensional cross-section of magnetic force provided by an embodiment of the present invention for a three-dimensional modeling and visualization method based on gravity and magnetic data; Figure 6 This is a schematic diagram of gravity isosurfaces, illustrating a three-dimensional modeling and visualization method based on gravity and magnetic data provided in an embodiment of the present invention. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] This invention discloses a 3D modeling and visualization method based on gravity and magnetic data. First, gravity and magnetic anomaly data are collected using instruments to obtain gravity and magnetic anomaly data. Then, gravity and magnetic field components at different depths are extracted from the gravity and magnetic anomaly data. Based on the geographical distribution of the collected gravity and magnetic anomaly data, a 3D mesh model is constructed. The density and magnetic susceptibility parameters corresponding to each grid cell in the 3D mesh model are used as model variables to be inverted, constructing an initial model. Gravity and magnetic responses in the initial model are calculated using gravity and magnetic field forward modeling operators, and compared with the collected gravity and magnetic anomaly data to obtain gravity and magnetic field residuals. Different weights are assigned to the gravity and magnetic field components at different depths, and the gravity and magnetic field residuals are mapped to corresponding grid cells with different weights, enabling the forward modeling of the 3D mesh model to obtain accurate gravity and magnetic anomaly data. Then, a comprehensive inversion objective function is constructed, and the 3D mesh model is inverted using the comprehensive inversion objective function to obtain a subsurface density distribution model and a magnetic susceptibility distribution model, indirectly inferring subsurface geological properties that cannot be directly measured, and achieving 3D visualization.

[0018] like Figure 1 and Figure 2 As shown, the present invention provides a three-dimensional modeling and visualization method based on gravity and magnetic data, which mainly includes the following steps: acquiring surface observation data of the study area, including gravity anomaly data and magnetic anomaly data, wherein the gravity anomaly data is the observation result of the density distribution of the subsurface medium after gravity forward modeling response, and the magnetic anomaly data is the observation result of the magnetic susceptibility distribution of the subsurface medium after magnetic field forward modeling response; performing preprocessing of zero-point drift correction and format conversion on the gravity anomaly data, and performing preprocessing of denoising and polarization on the magnetic anomaly data; Then, the preprocessed gravity and magnetic anomaly data are subjected to multi-scale decomposition to weaken the volume field superposition effect and obtain gravity and magnetic field components reflecting different depths underground. Multi-scale decomposition includes extracting low-frequency components of deep underground regions through up-extension processing, extracting local anomaly components of shallow underground regions through high-pass filtering, and extracting anomaly boundaries that enhance signals. Based on the spatial distribution of surface observation data, a three-dimensional grid model of the study area is constructed, and the underground space is divided into multiple continuous grid units. The density parameter and magnetic susceptibility parameter of each grid unit are used as model variables to be inverted. Initial values ​​are assigned to the model variables to construct an initial model. The gravity response and magnetic field response corresponding to the initial model are calculated by gravity forward modeling operator and magnetic field forward modeling operator, and compared with the observed gravity anomaly data and magnetic anomaly data to obtain the corresponding gravity residual and magnetic field residual; Finally, a comprehensive inversion objective function for gravity and magnetic fields is constructed. The comprehensive inversion objective function includes gravity data fitting terms, magnetic data fitting terms, and model regularization constraint terms. Weight parameters for different fitting terms are set using gravity and magnetic field components. Based on gravity residuals and magnetic field residuals, the conjugate gradient method is used to iteratively solve the comprehensive inversion objective function to obtain the underground density distribution model and magnetic susceptibility distribution model, which are then visualized in three dimensions.

[0019] Specifically, surface observation data of the study area is obtained. Multiple surface test points with different geographical locations are set up in the study area. Gravity anomaly data of the surface observation data is collected at the surface test points using a ground gravimeter, and magnetic anomaly data of the surface observation data is obtained at the surface test points using a magnetometer.

[0020] In this embodiment, the Dalad Banner detection area was selected as the target area, with the region ranging from 115.0° to 116.0° east longitude and 36.0° to 37.0° north latitude as the study area. This area is located in the 20th zone of the Gauss-Krüger 6° sub-zone of the Beijing 54 coordinate system. 500 surface test points were evenly distributed within the study area. For example, the coordinates of surface test point 1 (X=20623500m, Y=3650000m, H=126.5m) and surface test point 2 (X=20624000m, Y=3650000m, H=127.2m) were respectively. At each surface test point, gravity anomaly data were collected using a high-precision ground gravimeter to obtain the observation results of the underground medium density distribution after gravity response. Simultaneously, a proton magnetometer was used to collect the magnetic field response of the underground medium magnetic susceptibility distribution to obtain magnetic anomaly data.

[0021] Furthermore, the gravity anomaly data undergoes preprocessing including zero-point drift correction and format conversion, while the magnetic anomaly data undergoes preprocessing including denoising and polarization.

[0022] In this embodiment, the gravimeter will experience zero-point drift during the observation process due to the instrument's own stability. The gravimeter was zero-point calibrated before and after the observation at the test point, and the drift value was recorded as 0.59 mGal (the observation time was 4 hours, and the drift amount was linearly distributed). This is the original gravity anomaly data. The zero-point drift value is given by the zero-point drift correction formula: ; in, For corrected gravity anomaly data; The corrected gravity anomaly data is converted into an industry-standard format: X-coordinate-Y-elevation-corrected gravity value. After format conversion, invalid characters are simultaneously removed to ensure the data can be recognized by subsequent volume effect correction procedures. The final result of gravity anomaly data preprocessing is only the corrected gravity value.

[0023] The wavelet denoising algorithm is used to remove the influence of electromagnetic interference around the original magnetic anomaly data. The wavelet algorithm decomposes the original magnetic anomaly data to obtain the interference component. The interference component is then subtracted from the original magnetic anomaly data to obtain the denoised magnetic anomaly data.

[0024] The surface testing point is located in the Northern Hemisphere, where the geomagnetic field has an inclination. Therefore, the denoised magnetic anomaly data needs to be reoriented to the magnetic north pole to eliminate anomalous displacement caused by the magnetic inclination, ensuring the anomaly center corresponds to the center of the underground magnetic body. This polarization processing employs a sinusoidal correction method. The polarization coefficient is set according to the direction of the Northern Hemisphere magnetic field, and the product of the denoised magnetic anomaly data and the polarization coefficient is calculated to obtain the polarized magnetic anomaly data. This data is then preprocessed.

[0025] Furthermore, the preprocessed gravity and magnetic anomaly data are subjected to multi-scale decomposition to weaken the volume field superposition effect and obtain gravity and magnetic field components reflecting different depths underground. The multi-scale decomposition includes extracting low-frequency components in deep underground regions through up-extension processing, extracting local anomaly components in shallow underground regions through high-pass filtering, and extracting anomaly boundaries that enhance the signal.

[0026] The preprocessed gravity and magnetic anomaly data are subjected to multi-scale decomposition, specifically as follows: The gravity and magnetic anomaly data are gridded to construct a regular spatial data field. Based on this gridded data, the gravity and magnetic anomaly data are extended at different heights using an up-extension method to obtain low-frequency field components corresponding to different extension heights, which are used to characterize the structural information of deep underground layers. The gravity and magnetic anomaly data are then subjected to high-pass filtering to separate high-frequency anomaly components, which are used to characterize local anomaly information in shallow underground layers. The vertical derivative and analytical signal of the gravity and magnetic anomaly data are calculated to enhance the anomaly boundary response and extract gradient features reflecting the boundary of the geological body.

[0027] In this embodiment, the spatial range of the study area is determined according to the different geographical locations of the gravity and magnetic anomaly data, a uniform grid spacing is set, grid nodes with regular row and column distribution are generated, a regular spatial grid skeleton is constructed, and discrete gravity and magnetic anomaly data are interpolated and distributed to each grid node to obtain a regular spatial data field.

[0028] Based on the regular spatial data field, the upextension method is used to perform low-pass filtering on gravity and magnetic anomaly data. Different extension heights are set according to the height of the collected gravity and magnetic anomaly data. For gravity and magnetic anomaly data with higher elevations in the collected surface observation data, a higher extension height is set. An upextension operator with successively decaying values ​​is constructed. The gravity and magnetic anomaly data are multiplied by the corresponding decaying upextension operator to obtain the field components that suppress deep low frequencies and the field components that highlight deep low frequencies. The low frequency field components with different extension heights are used to characterize the structural information of deep underground fault rocks or deep sedimentary rocks.

[0029] Gravity and magnetic anomaly data are processed by high-pass filtering using Fourier transform to convert the gravity and magnetic anomaly data to the frequency domain. The low-pass filtered component consisting of the upward extension operator is subtracted from the frequency domain to obtain the remaining high-pass filtered component. Then, the high-pass filtered component is returned to the spatial domain by inverse Fourier transform to obtain discrete high-frequency anomaly components, which characterize the anomaly information of shallow local fault layers and fracture zones.

[0030] The gravity and magnetic anomaly data are transformed from the spatial domain to the frequency domain. The first-order vertical derivative of the gravity and magnetic anomaly data is calculated in the frequency domain. A Fourier transform is then used to transform the data from the spatial domain to the frequency domain, and the differentials of the gravity and magnetic anomaly data in the x and y directions are calculated as horizontal and vertical derivative operators. The data is then multiplied by these horizontal and vertical derivative operators, and an inverse Fourier transform is used to obtain the horizontal and vertical gravity and magnetic fields. This process converts the maxima in the gravity and magnetic anomaly data into zero-value lines, highlighting the boundaries of geological bodies at medium depths. Finally, a Fourier transform is used to calculate the second-order vertical derivative of the gravity and magnetic anomaly data, transforming the data from the spatial domain to the frequency domain. A second-order derivative is then performed on the data to obtain zero-value lines with sharper boundary responses, highlighting the boundaries of shallow geological bodies.

[0031] After calculating the vertical derivative of the gravity and magnetic anomaly data, the square root of the sum of the squares of the vertical derivatives is used to obtain the amplitude of the analytical signal. Gradient features of boundary responses of different geological bodies are extracted based on analytical signals with different amplitudes. The amplitude of the analytical signal exhibits significant maxima at locations of abrupt changes in physical properties, such as geological body boundaries and fault contact zones, while the amplitude is lower within homogeneous geological bodies. A threshold screening method is applied to the analytical signal amplitude. Analytical signals with amplitudes exceeding the threshold are selected as high-gradient abrupt change regions, while those within the threshold are selected as gentle background regions. This enhances the response of anomaly boundaries and accurately extracts the gradient features of geological body boundaries, fault structures, and lithological contact zones.

[0032] Furthermore, based on the spatial distribution of surface observation data, a three-dimensional grid model of the study area is constructed, dividing the underground space into multiple continuous grid units. The density parameter and magnetic susceptibility parameter of each grid unit are used as model variables to be inverted. Initial values ​​are assigned to the model variables to construct an initial model. When constructing a three-dimensional mesh model of the study area, the boundary range of the underground modeling space is first determined based on the spatial distribution range of the surface observation data, including the horizontal range and the vertical depth range. After determining the spatial range, the underground space is divided into multiple three-dimensional grid units according to the preset spatial resolution. Corresponding physical property parameters are defined for each grid unit. The density parameter and magnetic susceptibility parameter of each grid unit are set as model variables to be solved by inversion, which are used to characterize the physical property distribution characteristics of different underground locations.

[0033] After obtaining the initial model, forward modeling calculations of the gravity field and magnetic field are performed based on the three-dimensional mesh model and model variables. For each observation point on the surface, the gravity response of all underground mesh cells to the observation point is calculated, and the contributions of all mesh cells are accumulated to obtain the predicted gravity value corresponding to the observation point. Based on the magnetic susceptibility parameters of each grid cell, the magnetic field response generated at each observation point is calculated and accumulated to obtain the predicted magnetic field value corresponding to each observation point; thus, the predicted gravity data and predicted magnetic field data corresponding to each observation point are obtained.

[0034] In this embodiment, five spatial resolutions are preset, dividing the three-dimensional mesh cells into five types of regions. Different density and magnetic susceptibility parameters are defined for each mesh cell to perform forward modeling on the three-dimensional mesh model. For gravity anomaly data collected from each surface observation point, the geographical location is mapped to the constructed three-dimensional mesh model to obtain the gravity response of that observation point to the mesh cell. The gravity response contributions of all mesh cells are superimposed to calculate the predicted gravity value for that surface observation point. For magnetic anomaly data collected from each surface observation point, the geographical location is mapped to the constructed three-dimensional mesh model to obtain the magnetic response of that surface observation point to the mesh cell. The magnetic response contributions of all mesh cells are superimposed to obtain the predicted magnetic value for that surface observation point. Thus, the predicted gravity data and predicted magnetic field data corresponding to each observation point are obtained.

[0035] After obtaining the predicted gravity data and predicted magnetic field data, they are compared with the observed gravity anomaly data and magnetic anomaly data on a grid-by-grid basis, and the difference at each observation point is calculated. The difference between the predicted gravity data and the observed gravity anomaly data is taken as the gravity residual, and the difference between the predicted magnetic field data and the observed magnetic anomaly data is taken as the magnetic field residual.

[0036] In this embodiment, predicted gravity and magnetic field data are obtained through forward modeling of a 3D mesh model. These predicted gravity and magnetic field data are then compared with actual observed gravity and magnetic anomaly data. The difference between each predicted gravity data point and its corresponding observed gravity anomaly data is taken as the gravity residual, and the difference between each predicted magnetic field data point and its corresponding observed magnetic anomaly data is taken as the magnetic field residual. These gravity and magnetic field residuals reflect the differences between the current 3D mesh model and the actual surface observation data, providing accurate input sources for the subsequent comprehensive inversion objective function and improving the convergence speed and accuracy of the 3D modeling.

[0037] Subsequently, a comprehensive inversion objective function for gravity and magnetic fields was constructed. The comprehensive inversion objective function includes gravity data fitting terms, magnetic data fitting terms, and model regularization constraint terms. Weight parameters for different fitting terms were set using gravity and magnetic field components. The weight parameters were set larger as the depth of the gravity and magnetic field components increased. Based on the gravity residual and magnetic field residual, the conjugate gradient method was used to iteratively solve the comprehensive inversion objective function to obtain the underground density distribution model and the magnetic susceptibility distribution model, which were then visualized in three dimensions.

[0038] Gravity and magnetic data inversion is performed based on a 3D mesh model. A comprehensive inversion objective function is constructed, integrating gravity data fitting terms, magnetic data fitting terms, and model regularization constraint terms. The magnetic data fitting term is generated through a nonlinear equation, with the specific formula as follows: ; in, This represents the total number of magnetic observation locations. For the magnetic anomaly data collected at the i-th observation location in the study area, The predicted magnetic field data is obtained from the forward modeling of the three-dimensional mesh model at the i-th observation location. The weighting parameters for magnetic anomaly data. It is a constant.

[0039] The formula for the gravity data fitting term is: ; in, This represents the total number of gravity observation locations. For gravity anomaly data collected at the i-th observation location in the study area, The predicted gravity data is obtained from the forward modeling of the 3D mesh model at the i-th observation location. The weighting parameters for gravity anomaly data. It is a constant.

[0040] The formula for the model regularization constraint term is: = ; in, The model smoothing regularization constraint term is used. A smaller value for this constraint term results in smaller gradients in the model density and magnetic susceptibility, leading to better overall smoothness and clearer continuous gradient characteristics in the 3D mesh model. Q represents the total number of mesh elements, and j represents the j-th mesh element. This is the sum of all 3D meshes in the model; Let J be the density gradient of the j-th grid cell in the x-direction of the model. Let J be the density gradient of the j-th grid cell in the y-direction of the model. Let be the density gradient of the j-th grid cell in the z-direction of the model.

[0041] A comprehensive inversion objective function is constructed by fitting gravity data, magnetic data, and model regularization constraints. This comprehensive inversion objective function is as follows: ; in, For the gravity data fitting term, This is a term for fitting magnetic force data. This is the model regularization constraint term.

[0042] Furthermore, the comprehensive inversion objective function performs joint inversion of physical properties for the spatial characteristics of gravity and magnetic fields at different scales. Based on the gravity residual and magnetic field residual, the conjugate gradient method is used to iteratively solve the comprehensive inversion objective function. The gravity residual and magnetic field residual are fused and added to the comprehensive inversion objective function. The minimum solution of the comprehensive inversion objective function is iterated. In each iteration, the physical property gradient vector of the comprehensive inversion objective function with respect to density and magnetic susceptibility is calculated. The conjugate direction is determined by the gradient data of the previous iteration. The search is carried out along the conjugate direction to determine the iteration step size, so as to obtain the density distribution model and magnetic susceptibility distribution model of the study area and perform three-dimensional visualization.

[0043] In this embodiment, the comprehensive inversion objective function performs joint inversion of density and magnetic susceptibility of low-frequency field components and high-frequency anomaly components at different scales. After matching the gravity residuals and magnetic field residuals at the same test point with the actual measured gravity anomaly data and magnetic anomaly data, the comprehensive inversion objective function is iteratively minimized using the conjugate gradient method.

[0044] First, it is deduced whether the predicted gravity and magnetic field data correspond to the actual measured gravity and magnetic anomaly data. Then, the gradient vectors of density and magnetic susceptibility of the predicted gravity and magnetic field data are iteratively calculated. In each iteration, an iterative gradient is set, and the predicted gravity data is iteratively compared with the actual measured gravity anomaly data, thus ensuring that the predicted magnetic field data corresponds with the actual measured magnetic anomaly data. Next, iterative conditions for the predicted gravity data are set, including reaching gravity anomaly data or satisfying a certain number of iterations. A density gradient is added to the predicted gravity data to iterate over the actual gravity anomaly data, resulting in a correlation between the predicted gravity data and the actual gravity anomaly data. After the gravity anomaly data are equalized, the iteration direction is set. At the test point, the obtained predicted gravity data is used to rotate the iteration vector direction at that point to find the conjugate direction pointing to the next test point of predicted gravity data. The search is performed along the conjugate direction, and the gradient of the iteration step size is set to minimize the value of the comprehensive inversion objective function corresponding to the density of the predicted gravity data. The predicted gravity data of the next test point is searched step by step along the conjugate direction with a gradually increasing gradient, and then the predicted gravity data of all test points are obtained. Finally, the predicted gravity data of all test points are used to construct a density distribution model to infer the properties of the underground geology.

[0045] When iteratively minimizing the magnetic susceptibility of the comprehensive inversion objective function using the conjugate gradient method, the predicted magnetic field data corresponds to the actual measured magnetic anomaly data. Iteration conditions for the predicted magnetic field data are then set, including reaching the magnetic anomaly data or satisfying the required number of iterations. Magnetization values ​​are added to the predicted magnetic field data to iterate over the actual magnetic anomaly data until the predicted magnetic field data equals the actual magnetic anomaly data. An iteration direction is then set. Using the obtained predicted magnetic field data, the iteration vector direction is rotated at that point to find the conjugate direction pointing to the next predicted magnetic field data test point. The search continues along the conjugate direction, with a gradient for the iteration step size. The magnetic susceptibility of the next predicted magnetic field data is searched by gradually increasing the gradient along the conjugate direction. The value of the comprehensive inversion objective function corresponding to the magnetic susceptibility of the next predicted magnetic field data is minimized, thus obtaining the predicted magnetic field data for the next test point. This process is repeated for all test points. Finally, the predicted magnetic field data from all test points are used to construct a magnetic susceptibility distribution model to obtain the properties of the underground geology, resulting in a 3D visualization.

[0046] Density distribution models effectively visualize density gradient zones within the mantle, revealing spatial differences in the density of subsurface media and providing a clear three-dimensional spatial distribution of different lithological strata, such as fault rocks and sedimentary rocks. Simultaneously, magnetic susceptibility distribution models accurately identify anomalous zones where the magnetic susceptibility of the subsurface media undergoes significant abrupt changes, capturing information on local geological structural changes and clearly displaying geological anomalies caused by local fault zones and fracture zones. Through density and magnetic susceptibility distribution models, the spatial distribution of active faults within the Earth's crust and the range of high-magnetization accumulation zones can be precisely characterized. Based on this, the spatial location of potential strong earthquake source bodies can be predicted, achieving a clear and visual representation of potential strong earthquake hazard zones. This provides crucial data support for early earthquake monitoring and warning, enhancing earthquake disaster prevention and mitigation capabilities.

[0047] like Figure 3 As shown, this invention provides a density distribution model for gravity three-dimensional modeling in a method for three-dimensional modeling and visualization based on gravity and magnetic data. The density distribution model visualizes gravity anomaly data in the Dalad Banner exploration area. In the density distribution model, warm colors represent basement uplift or high-gradient fault rocks, while cold colors represent granitic lighter rock masses such as sedimentary basin abrupt change zones or fracture zones.

[0048] like Figure 4 As shown, this invention demonstrates a magnetic susceptibility distribution model for three-dimensional magnetic modeling. In the magnetic susceptibility distribution model, warm colors represent highly magnetized geological bodies, such as igneous rocks or strongly magnetic intrusive bodies; cool colors represent weakly magnetic or non-magnetic geological bodies, mostly sedimentary rocks, acidic rocks, or strongly altered demagnetized areas. In this invention, the warm and cool colors in the density distribution model and magnetic susceptibility distribution model are clearly distinguishable, indicating that the density distribution model and magnetic susceptibility distribution model constructed using gravity and magnetic anomaly data are highly accurate and can accurately predict the geological properties of the study area, thereby improving earthquake disaster prevention and mitigation capabilities.

[0049] like Figure 5 As shown, the figure represents a three-dimensional magnetic cross-section in the study area, illustrating abrupt changes in magnetic susceptibility. The figure also shows the different lithologies on both sides of the fault in the study area, which helps to identify surface slip rate and seismic activity in active faults and predict the geological properties of earthquakes and volcanoes.

[0050] like Figure 6 As shown, the gravity isosurface of the Dalad Banner exploration area shows the density distribution and whether the crust is in a state of gravity equilibrium. The upward convexity of the isosurface indicates the basement uplift, volcanic rocks, or areas of crustal thinning, while the downward depression of the isosurface indicates sedimentary basins or magmatic layer areas. The gravity isosurface can be used to monitor the dynamic deformation of the crust and is of great significance in geodesy and seismic geology research.

[0051] This invention discloses a 3D modeling and visualization method based on gravity and magnetic data. It acquires gravity and magnetic anomaly data from the study area, decomposes the gravity and magnetic anomaly data into multiple components with different depth sensitivities, and incorporates these residuals into a constructed comprehensive inversion objective function through forward modeling and calculation of gravity and magnetic field residuals. Differential weight constraints are introduced during the inversion process to effectively separate the low-frequency interference background field of deep underground space from local anomalies of shallow underground space. This reduces the impact of superposition effects on underground exploration modeling, improves the resolution and stability of underground physical property distribution inversion, and effectively predicts underground geological properties.

[0052] It should be noted that the numerous details included in the above description are merely illustrative of the invention and not intended to limit it. In other embodiments of the invention, the method may have more, fewer, or different steps, and the order, inclusion, function, etc., of the steps may differ from those described and illustrated.

Claims

1. A three-dimensional modeling and visualization method based on gravity and magnetic data, characterized in that, The method includes the following steps: Surface observation data of the study area are obtained, including gravity anomaly data and magnetic anomaly data. The gravity anomaly data is the observation result of the density distribution of the subsurface medium after gravity forward modeling response, and the magnetic anomaly data is the observation result of the magnetic susceptibility distribution of the subsurface medium after magnetic field forward modeling response. The gravity anomaly data is preprocessed with zero-point drift correction and format conversion, and the magnetic anomaly data is preprocessed with denoising and polarization. The preprocessed gravity and magnetic anomaly data are subjected to multi-scale decomposition to weaken the volume field superposition effect and obtain gravity and magnetic field components reflecting different depths underground. The multi-scale decomposition includes extracting low-frequency components of deep underground regions through upward extension processing, extracting local anomaly components of shallow underground regions through high-pass filtering, and extracting anomaly boundaries that enhance the signal. Based on the spatial distribution of surface observation data, a three-dimensional grid model of the study area is constructed, the underground space is divided into multiple continuous grid units, and the density parameter and magnetic susceptibility parameter of each grid unit are used as model variables to be inverted. Initial values ​​are assigned to the model variables to construct an initial model. The gravity response and magnetic field response corresponding to the initial model are calculated by gravity forward modeling operator and magnetic field forward modeling operator, and compared with the observed gravity anomaly data and magnetic anomaly data to obtain the corresponding gravity residual and magnetic field residual; A comprehensive inversion objective function for gravity and magnetic fields is constructed, which includes gravity data fitting terms, magnetic data fitting terms, and model regularization constraint terms. Weight parameters for different fitting terms are set using the gravity and magnetic field components. Based on the gravity residuals and magnetic field residuals, the comprehensive inversion objective function is iteratively solved using the conjugate gradient method to obtain the underground density distribution model and the magnetic susceptibility distribution model, which are then visualized in three dimensions.

2. The method for three-dimensional modeling and visualization based on gravity and magnetic data according to claim 1, characterized in that, Surface observation data of the study area is obtained. Multiple surface test points with different geographical locations are preset in the study area. Gravity anomaly data of the surface observation data is collected at the surface test points using a ground gravimeter. Magnetic anomaly data of the surface observation data is collected at the surface test points using a magnetometer.

3. The method for three-dimensional modeling and visualization based on gravity and magnetic data according to claim 1, characterized in that, The preprocessed gravity anomaly data and magnetic anomaly data are subjected to multi-scale decomposition processing, including the following steps: The gravity and magnetic anomaly data are gridded to construct a regular spatial data field; Based on the gridded data, the gravity anomaly data and magnetic anomaly data are extended at different heights using the up-extension method to obtain low-frequency field components corresponding to different extension heights, which are used to characterize the structural information of deep underground layers. High-pass filtering was performed on gravity anomaly data and magnetic anomaly data to separate high-frequency anomaly components, which were used to characterize local anomaly information in shallow underground layers. Vertical derivatives and analytical signals are calculated for gravity anomaly data and magnetic anomaly data to enhance the response of anomaly boundaries and extract gradient features that reflect the boundaries of geological bodies.

4. The method for three-dimensional modeling and visualization based on gravity and magnetic data according to claim 1, characterized in that, When constructing a three-dimensional mesh model of the study area, the boundary range of the underground modeling space is first determined based on the spatial distribution range of the surface observation data, including the horizontal range and the vertical depth range. After determining the spatial range, the underground space is divided into multiple three-dimensional grid units according to the preset spatial resolution. Corresponding physical property parameters are defined for each grid unit. The density parameter and magnetic susceptibility parameter of each grid unit are set as model variables to be solved by inversion, which are used to characterize the physical property distribution characteristics of different underground locations.

5. The method for three-dimensional modeling and visualization based on gravity and magnetic data according to claim 1, characterized in that, After obtaining the initial model, based on the three-dimensional mesh model and model variables, forward modeling calculations of the gravitational and magnetic fields are performed, and the following steps are executed: For each observation point on the surface, the gravity response of all underground grid cells to that observation point is calculated, and the contributions of all grid cells are summed to obtain the predicted gravity value corresponding to that observation point. Based on the magnetic susceptibility parameters of each grid cell, the magnetic field response generated at each observation point is calculated and accumulated to obtain the predicted magnetic field value corresponding to each observation point. Obtain the predicted gravity data and predicted magnetic field data corresponding to each observation point.

6. The method for three-dimensional modeling and visualization based on gravity and magnetic data according to claim 5, characterized in that, After obtaining the predicted gravity data and predicted magnetic field data, they are compared with the observed gravity anomaly data and magnetic anomaly data on a grid-by-grid basis, and the difference at each observation point is calculated. The difference between the predicted gravity data and the observed gravity anomaly data is taken as the gravity residual, and the difference between the predicted magnetic field data and the observed magnetic anomaly data is taken as the magnetic field residual.

7. The method for three-dimensional modeling and visualization based on gravity and magnetic data according to claim 1, characterized in that, Gravity and magnetic data inversion is performed based on the aforementioned 3D mesh model. A comprehensive inversion objective function is constructed, integrating gravity data fitting terms, magnetic data fitting terms, and model regularization constraint terms. The magnetic data fitting term is generated through a nonlinear equation, specifically as follows: ； in, This represents the total number of magnetic observation locations. For the magnetic anomaly data collected at the i-th observation location in the study area, The predicted magnetic field data is obtained from the forward modeling of the three-dimensional mesh model at the i-th observation location. The weighting parameters for magnetic anomaly data. It is a constant.

8. The method for three-dimensional modeling and visualization based on gravity and magnetic data according to claim 7, characterized in that, A comprehensive inversion objective function is constructed that integrates the gravity data fitting term, the magnetic data fitting term, and the model regularization constraint term. The comprehensive inversion objective function is as follows: ； in, For the gravity data fitting term, For the magnetic data fitting term, For the physical property constraints of the model.

9. A method for three-dimensional modeling and visualization based on gravity and magnetic data according to claim 8, characterized in that, The comprehensive inversion objective function performs joint inversion of physical properties of gravity and magnetic spatial characteristics at different scales. The weight parameters of different fitting terms are set using the gravity and magnetic field components. Based on the gravity residual and magnetic field residual, the comprehensive inversion objective function is iteratively minimized using the conjugate gradient method. In each iteration, the gradient vector of the comprehensive inversion objective function with respect to density and magnetic susceptibility is calculated. The conjugate direction is determined by the gradient data of the previous iteration. The search is performed along the conjugate direction to determine the iteration step size, thereby obtaining the density distribution model and magnetic susceptibility distribution model of the study area.

Images