Source body positioning method based on intelligent gravity and magnetic field separation and Euler deconvolution

Abstract

The invention belongs to the technical field of gravity and magnetic method exploration data processing, and relates to a source body positioning method based on intelligent gravity and magnetic field separation and Euler deconvolution, which comprises the following steps: setting gravity and magnetic abnormal field modeling parameters according to geological and physical property data of a research area; generating an initial seed point in the underground three-dimensional grid and expanding and growing the initial seed point to generate Q independent abnormal field models; mapping the position of the model to a real coordinate, and calculating to obtain a forward modeling abnormal field data set; extracting the maximum value of the global absolute value of the data set and sorting in an ascending order, superposing the first p absolute values to generate a composite field, and constructing a field separation data set by taking the p absolute value as a target field and the composite field; using the field separation data set to train the field separation network until convergence; inputting measured data score departure field and residual field data, and selecting data for iterative separation according to requirements to obtain a field separation result; and obtaining a preliminary positioning result by using a field separation result and combining with an Euler deconvolution method, and determining final source body positioning after screening. The method has the beneficial effects that multi-scale geological target field separation can be realized, and source body positioning can be determined.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of gravity and magnetic exploration data processing, and particularly relates to a source body positioning method based on intelligent gravity and magnetic field separation and Euler deconvolution. BACKGROUND

[0002] The purpose of source body positioning of gravity and magnetic data is to obtain the distribution position, structural form and other information of target geological bodies according to the gravity or magnetic method data obtained in the field. Among them, Euler deconvolution is an automatic calculation method that can determine the position and depth of geological bodies by using gravity and magnetic grid data. It can be used to calculate the burial depth and position of geological bodies and is widely used in various resource exploration work. However, due to the complex mineral alteration types and variable geological structures in the deep crust, the obtained gravity and magnetic anomaly field parameters usually show superimposed effects of different target geological body anomalies. This makes it difficult for the Euler deconvolution method to accurately determine the boundary range and shape of the underground ore body, and reduces the accuracy of gravity and magnetic anomaly field interpretation and analysis.

[0003] From the traditional sense, the field separation of gravity and magnetic exploration data is mainly to divide the residual gravity and magnetic anomaly and the regional gravity and magnetic anomaly. However, with the deepening of related research, the accuracy of field separation needs to be improved. Therefore, the multi-scale series method is proposed, that is, in addition to the regional gravity and magnetic anomaly, the anomaly information generated by different underground geological bodies also needs to be separated. Multi-scale weak anomalies are relative, for example: two geological bodies with similar scales but different burial depths, the anomaly generated by the geological body with deeper burial depth can be regarded as a weak anomaly; and two geological bodies with the same burial depth but different scales, the anomaly generated by the geological body with smaller scale is regarded as a weak anomaly.

[0004] Traditional separation methods such as analytic continuation method, wavelet multi-scale analysis method and polynomial fitting method mainly study regional field and local field, and few of them study residual anomaly. The traditional methods that can separate fields and extract weak anomalies such as wavelet multi-scale analysis, blind source separation and empirical mode decomposition technology depend on the threshold setting involved in the experiment, and the threshold setting is affected by subjective factors. Therefore, the potential value of residual field separation and residual anomaly extraction still needs further research. SUMMARY

[0005] Technical problems to be solved In view of the above-mentioned shortcomings and deficiencies of the prior art, the present application provides a source body positioning method based on intelligent gravity and magnetic field separation and Euler deconvolution, which solves the technical problems that the traditional separation method mainly studies regional field and local field and few of them study residual anomaly, and the method that can separate fields and extract weak anomalies depends on the threshold setting affected by subjective factors.

[0006] Technical scheme In order to achieve the above object, the main technical scheme adopted by the present application comprises: The present application provides a source body positioning method based on intelligent gravity and magnetic field separation and Euler deconvolution, comprising the following steps: Step 1: According to the geological and physical data of the study area, set the gravity and magnetic anomaly field modeling parameters; Step 2: Randomly generate a plurality of initial seed points in the three-dimensional grid space of the study area, and expand and grow in a random direction according to a predetermined rule until Q random three-dimensional models of independent anomaly fields are generated; Step 3: Map the random three-dimensional model position to the real physical coordinates, and calculate each model to obtain the forward anomaly field data set; Step 4: Extract the maximum value of the global absolute value of the forward anomaly field data set, and sort all data in ascending order to obtain an ordered model sequence ; Step 5: Extract the first p data of the ordered model sequence to generate a composite field , take the data as the separation target field , and form a field separation data set ; Step 6: Train the field separation network with the field separation data set until the network reaches the preset convergence standard; Step 7: Input the collected measured gravity and magnetic exploration data into the trained field separation network, output the separation field data, and calculate the difference between the measured gravity and magnetic exploration data and the separation field data to obtain the residual field data; Step 8: According to the geological interpretation requirements, select the separation field data or the residual field data as the new input, input the trained field separation network for iterative separation, and obtain the field separation result; Step 9: According to the field separation result, use the joint Euler deconvolution method to solve and obtain the preliminary source body positioning result; Step 10: Screen the preliminary source body positioning result to finally determine the source body positioning result.

[0007] As a further improvement of the present application, the step 1 is specifically: based on the geological mapping data and physical property database of the study area, the density and magnetic susceptibility range of different lithology zones are counted, the physical property parameter interval, the initial seed size and the maximum growth volume constraint of the anomaly field modeling are set; the observation plane position and grid resolution are defined; The anomaly field modeling parameters include gravity gradient data, gravity anomaly data, magnetic anomaly data and magnetic gradient data.

[0008] As a further improvement of the present application, the step 2 of generating Q random three-dimensional models of independent anomaly fields comprises the following steps: Step 2.1: Establish a growth path stack for each initial seed, and record the current growth coordinates; Step 2.2: Determine whether the current growth coordinates of the initial seed meet the growth conditions; Specifically, if the neighborhood grid within the three-dimensional grid space under the initial seed is not occupied and is within the preset growth space, it is marked as a growable area; if the growth is blocked, return to the last growth node and reselect the growth direction; Step 2.3: Expand the seed volume in a random direction within the legal growth space until the preset growth number or volume limit is reached; Step 2.4: Loop the above steps until Q independent random three-dimensional models of abnormal fields are generated.

[0009] As a further improvement of the present application, the model obtained by growing the random seed in the three-dimensional grid in step 3 is converted to the real physical space position, and is calculated according to the following formula: ; In the formula, and respectively represent the real physical space position and the grid position corresponding to the th grid point in the random three-dimensional model, represents the interval of the real space subdivision grid, represents the minimum value of the grid point coordinates in a certain direction, respectively represent different coordinate axes, represents the grid point coordinates in a certain direction.

[0010] As a further improvement of the present application, the gravity or magnetic forward calculation is performed on each random three-dimensional model in step 3 to obtain the corresponding forward abnormal field data.

[0011] As a further improvement of the present application, the field separation data set in step 5 is specifically composed of: taking the first forward abnormal field data in the ordered model sequence to generate a composite stacking field by stacking the first single separation target field as the target output, calculated according to the following formula: ; The composite stacking field and the separation target field constitute a field separation data set , and the training set and the validation set are divided according to the proportion.

[0012] As a further improvement of the present invention, step 6 utilizes a field separation dataset. The training ground separation network is specifically as follows: Construct a field separation network, establish a nonlinear mapping from the composite superimposed field to the target separation field through the field separation network, and optimize the weight parameters of the field separation network using the partitioned training set and validation set until the field separation network reaches the preset convergence criterion; The field separation network consists of a field separation estimation network and a field separation sub-network. The field separation estimation network consists of multiple convolutional layers and activation functions, used to extract multi-scale features of the composite superimposed field. The field separation sub-network consists of an encoder for feature extraction and data compression and a decoder for data reconstruction. The calculation formula for the convolutional layer of the field separation estimation network is as follows: ; ; ; In the formula, , Indicates the current convolutional layer Input data, This represents the output of the current convolutional layer after performing a convolution operation on the input data; This represents the weight values ​​of the current convolutional layer. Indicates the number of channels for the input data; , These represent the width and depth of the current convolutional layer, respectively. and This represents the stride and padding values ​​used in the convolution operation on the input data. This represents the bias coefficient of the current convolutional layer; Represents a convolutional layer Size, It can be represented as: ; In the formula, This represents the activation function applied during the current model training.

[0013] As a further improvement of the present invention, step 7 specifically involves: inputting the collected measured gravity and magnetic exploration data into the trained field separation network, outputting the separated field data, and calculating the difference between the measured gravity and magnetic exploration data and the separated field data as the remaining field data according to the following formula: ; In the formula, For the remaining field data, The collected measured gravity and magnetic exploration data, This is the data for the separation field.

[0014] As a further improvement of the present application, the step 9 is based on the field separation result, and a joint Euler deconvolution method is used for solving, specifically: A joint Euler deconvolution method for gravity and magnetic anomaly data is proposed by calculating the vertical derivative, and the formula principle is based on the Euler homogeneous equation: ; Wherein, And Respectively, the observation point and the center coordinates of the geological body, Is the construction index, And Background field and gravity and magnetic anomaly data respectively; The vector product form is: ; In order to reduce the selection of construction index N and the interference of background field B, the second order derivative in x, y and z direction is obtained, and the joint Euler deconvolution formula of gravity gradient data or magnetic gradient data is obtained: ; The vector product form is: ; The sliding window method is used to solve the equation, and each sliding window solves the corresponding , that is, the preliminary source body positioning result, Is the number of solutions.

[0015] The beneficial effects of the present application are: Different scales, different depths, different scale geological targets can be separated, and the source body positioning result can be finally determined; the method is applied to the theoretical model and the measured data, and the feasibility of the algorithm is verified.

[0016] It has the advantages of fast and accurate field separation, can effectively identify the boundary and shape information of underground abnormal body, is suitable for interpretation of measured gravity and magnetic data, and has important significance for improving the processing and interpretation accuracy of geophysical gravity and magnetic field data. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 A flowchart of a source body positioning method based on intelligent gravity and magnetic field separation and Euler deconvolution is provided for the embodiment of the present application; Figure 2 A model space position schematic diagram generated by the random generation method in the present embodiment is shown; Figure 3 A theoretical model space position and physical property diagram in the present embodiment is shown; Figure 4The superimposed observation data generated for model I-1 and model I-2 in the embodiment; Figure 5 The observation data generated for I-1 in the embodiment; Figure 6 The observation data generated for I-1 in the embodiment; Figure 7 The anomaly data of I-1 obtained by field separation in the embodiment; Figure 8 The anomaly data of I-2 obtained by field separation in the embodiment; Figure 9 The processed magnetic anomaly data in the embodiment; Figure 10 The field separation result of the real magnetic anomaly data in the embodiment; Figure 11 The source body result obtained by the Euler deconvolution method for the field separation result in the embodiment. DETAILED DESCRIPTION

[0018] In order to better explain the present application, facilitate understanding, the following will be combined with the drawings, through specific embodiments, the present application is described in detail.

[0019] In order to better understand the above technical solutions, the following will be described in more detail with reference to the exemplary embodiments of the present application. Although the exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided in order to enable a clearer, more thorough understanding of the present application, and to enable the scope of the present application to be fully conveyed to those skilled in the art.

[0020] In the embodiment, a source body positioning method based on intelligent gravity and magnetic field separation and Euler deconvolution, as shown in Figure 1 includes the following steps: Step 1: According to the geological and physical data of the study area, set the gravity and magnetic anomaly field modeling parameters; Specifically, based on the geological mapping data and physical property database of the study area, the density and magnetic susceptibility range of different lithology zones are counted, and the physical property parameter interval, initial seed size and maximum growth volume constraint of the gravity and magnetic anomaly field modeling are set; define the observation plane position and grid resolution; The gravity and magnetic anomaly field parameters include gravity gradient data, gravity anomaly data, magnetic anomaly data and magnetic gradient data.

[0021] In the embodiment, the parameter setting is performed for the area where the measured data used subsequently is located, and the remaining density generation interval is set to [-1.0, 1.0] g / cm 3, the remanent magnetization generation interval is set as [1, 100] A / m, the initial seed size is set as 50m*50m*50m, and the to-be-grown seed quantity generation interval is set as [2, 5].

[0022] Step 2: A plurality of initial seed points are randomly generated in the underground three-dimensional grid space of the study area, and are expanded in a random direction according to a preset rule until Q random three-dimensional models of independent anomaly fields are generated; Step 2.1: A growth path stack is established for each initial seed to record the current growth coordinates; Step 2.2: It is judged whether the current growth coordinates of the initial seed meet the growth condition; Specifically, if the neighborhood grid of the initial seed in the underground three-dimensional grid space is not occupied and is in the preset growth space, it is marked as a growable area; if the growth is blocked, the last growth node is returned and the growth direction is reselected; Step 2.3: The seed volume is expanded in a random direction in the legal growth space until the preset growth number or the volume upper limit is reached; Step 2.4: The above steps are repeatedly executed until Q (seed number*single seed growth number) random three-dimensional models of independent anomaly fields are generated.

[0023] In this embodiment, for the convenience of calculation, the continuous underground geometric space is discretized into a limited number of small units, that is, the finite element grid is used to divide the underground three-dimensional space of the study area into a 64*64*20 grid; the generated model schematic diagram is as shown in Figure 2 The seed number generation interval is [2, 5], and the single seed growth number interval is [16, 81].

[0024] Step 3: The position of the random three-dimensional model is mapped to the real physical coordinates, and the forward anomaly field data set is calculated for each model; Specifically, the model obtained by growing the random seed in the three-dimensional grid is converted to the real physical space position, and the following formula is used for calculation:

[0025] In the formula, and respectively represent the real physical space position and the grid position corresponding to the th grid point in the random three-dimensional model, represents the real space subdivision grid spacing, represents the minimum value of the grid point coordinates in a certain direction, respectively represent different directions, represents the grid point coordinates in a certain direction.

[0026] The virtual coordinates are used when generating the random three-dimensional model, and need to be converted into real physical coordinates before subsequent forward calculation; each random three-dimensional model is discretized by a grid, and a gravity or magnetic force forward algorithm based on a discrete integral equation is used to calculate the forward anomaly field data generated by the discretized model on the observation plane.

[0027] Step 4: Extract the maximum value of the global absolute value of the forward anomaly field data set, and sort all data in ascending order to obtain an ordered model sequence ; Specifically, the maximum value of the global absolute value of the anomaly field data is extracted, and all forward anomaly field data is sorted in ascending order according to the maximum value to generate an ordered model sequence .

[0028] In this embodiment, the forward calculation is performed in the Cartesian coordinate system, and the coordinate system has a distribution range of 0m to 3200m on the X axis, 0m to 3200m on the Y axis, and 0m to 1000m on the Z axis. The distance between the measuring point and the measuring line is 50m. The gravity and magnetic anomaly and gradient data of the randomly grown anomaly body model are calculated.

[0029] Step 5: Extract the first data in the ordered model sequence to generate a composite field , take the th data as a separation target field , and form a field separation data set ; Specifically, the first forward anomaly field data in the ordered model sequence is taken to generate a composite stacking field as the network input, and the th single separation target field is taken as the target output , which is calculated according to the following formula:

[0030] The composite stacking field and the separation target field form a field separation data set , which is divided into a training set and a verification set according to a preset ratio .

[0031] In this embodiment, 2x10 4 groups of data are generated for each anomaly field data, and the training set and the verification set are divided according to a ratio of 4:1.

[0032] Step 6: Use the field separation data set to train the field separation network until the network reaches a preset convergence standard. Specifically, first, a field separation network is constructed, a nonlinear mapping from the composite superimposed field to the target separated field is established through the field separation network, and the weight parameters of the field separation network are optimized using the divided training set and the validation set until the field separation network reaches the preset convergence standard. The field separation network is composed of a field separation estimation network and a field separation sub-network, the field separation estimation network is composed of multiple convolution layers and activation functions, and is used to extract multi-scale features of the composite superimposed field, and the field separation sub-network is composed of an encoder for feature extraction and data compression and a decoder for data reconstruction.

[0033] The calculation formula of the field separation estimation network convolution layer is as follows:

[0034]

[0035]

[0036] In the formula, , represents the input data of the current convolution layer , and represents the output result of the current convolution layer after convolution operation on the input data. represents the weight value of the current convolution layer, represents the channel number of the input data. , respectively represent the width and depth of the current convolution layer. and represent the step size and edge padding value of the convolution operation on the input data, represents the bias coefficient of the current convolution layer. represents the size of the convolution layer , which can be represented as:

[0037] In the formula, represents the activation function applied by the current training network.

[0038] In this embodiment, the field separation estimation network is composed of 5 convolution layers and 5 nonlinear activation functions ReLU layers, and is used to estimate and preliminarily extract features of the data in the field separation data set . Then, the estimation results and the preliminary features are input into the field separation sub-network. In the field separation sub-network, the encoder gradually reduces the data size through multiple downsampling operations, and then the high-dimensional features are input into the decoder, which gradually restores the data size through upsampling operations, and finally realizes the result output through the output layer of 1x1 convolution. The parameters used are: learning rate 1x10-3 The learning rate decay strategy uses an equal-interval adjustment of the learning rate, reducing it by half every 200 iterations. The batch size is 128, the number of iterations is 200, the mean squared error (MSE) is used as the loss function, and the adaptive moment estimation (Adam) is used as the optimizer. When the validation set loss is continuous... Training terminates when the number of iterations has not decreased or the target number of iterations has been reached, and the optimal weights are saved, thus completing the training of the field separation network.

[0039] Step 7: Input the collected measured gravity and magnetic exploration data into the trained field separation network, output the separated field data, and calculate the remaining field data based on the difference between the measured gravity and magnetic exploration data and the separated field data; Specifically, the collected measured gravity and magnetic exploration data are input into a trained field separation network, which outputs separated field data. The remaining field data is calculated as the difference between the measured data and the separated field data according to the following formula:

[0040] In the formula, For the remaining field data, The collected measured gravity and magnetic exploration data, This is the data for the separation field.

[0041] Step 8: Select the separated field data or residual field data as new input according to the geological interpretation requirements, input the trained field separation network for iterative separation, and obtain the field separation results; Specifically, based on the needs of geological interpretation, or As new input data, it is fed into the trained field separation network, and the process is repeated to achieve iterative separation of multi-scale anomalies, thereby obtaining field separation results. Then, for each separation result, the next step of source body localization is carried out.

[0042] The purpose of field separation mentioned here is to decompose the superimposed anomalies generated by multiple magnetic bodies into multiple independent magnetic anomaly fields. By calculating each individual anomaly field separately, the positioning accuracy can be improved.

[0043] In this embodiment, a gravity anomaly model different from the training and validation sets is established. Models I-1 and I-2 are obtained after 22 and 17 seed growth cycles, respectively, with densities of -0.24 g / cm³. 3 With 0.17 g / cm 3 The burial depths are 750m and 550m respectively. Their spatial locations are as follows: Figure 3 As shown. Calculate its gravity anomaly data V. z And the field separation network was used for testing, and the results are as follows: Figures 4-8The field separation results of the model I-1 and the model I-2 are shown in FIG. 3, and the mean square errors of the observed anomalies generated by the theoretical models I-1 and I-2 are shown in Table 1. From the image distribution, the field separation results of the model I-1 and the model I-2 are well corresponding to the observed data in the spatial position, the amplitude size and the anomaly distribution; then, the gravity gradient data of the model I is separated , , , , , and the gravity anomaly data are separated

[0044] Table 1: The mean square errors of the separation results of the model I and the real data

[0045] In order to verify the practicability of the method, the magnetic anomaly data of Macheng area in Hebei Province of China is used for testing, a test area with an area of about 3200m x 3200m is selected for the real data test, the underground is divided into 64 x 64 x 20 small cuboids, and each small cuboid has a size of 50 x 50 x 50m 3 . After the daily variation correction, the IGRF correction and the polarization processing, the processed magnetic anomaly data are shown in Figure 9 .

[0046] The field separation results are shown in Figure 10 , and the method proposed in the application is used to realize the main anomaly extraction of the magnetic anomaly data of Macheng area in Hebei Province.

[0047] Step 9: according to the field separation results, the joint Euler deconvolution method is used to solve, and the preliminary source body positioning results are obtained; Specifically, a joint Euler deconvolution method for gravity and magnetic anomaly data is proposed by calculating the vertical derivative, and the formula principle is based on the Euler homogeneous equation: (8)

[0048] wherein, and are the coordinates of the observation point and the center of the geological body respectively, is the construction index, and are the background field and the gravity and magnetic anomaly data respectively.

[0049] The formula is converted into a vector product form as follows: (9) In order to reduce the interference of the selection of the construction index N and the background field B, second-order derivatives are performed on the x, y and z directions, and a joint Euler deconvolution formula of the gravity gradient data or the magnetic gradient data is obtained: (10) The formula is converted into a vector product form as follows: (11) The sliding window method is used to solve the equation, and each sliding window is used to solve the corresponding , which is a preliminary source body positioning result, is the number of solutions.

[0050] Step 10: screening the preliminary source body positioning result to finally determine the source body positioning result; Specifically, for the solutions obtained by solving the Euler deconvolution equation, the effectiveness screening is used to obtain the final result, that is, the finally determined source body positioning result, and the specific effectiveness screening method includes but is not limited to the horizontal gradient filtering, the main body anomaly distance criterion and the dispersion index criterion.

[0051] In the embodiment, the coefficient of the horizontal gradient filtering method is set to 1, the dispersion index is set to 6, and the dispersion index criterion action radius is set to 1.5 times the line spacing.

[0052] The Euler deconvolution method is used for source body positioning for the field separation result, and the result is shown in Figure 11 , reflecting the profile and boundary of the underground anomaly body.

[0053] Through the above operation process, the source body positioning method based on the intelligent gravity and magnetic field separation and the Euler deconvolution method proposed in the application can perform rapid and accurate field separation, highlight the main anomaly field data, the joint Euler deconvolution method proposed can reflect the spatial position of the underground anomaly body, and is beneficial to improve the application of gravity and magnetic exploration in practical work.

[0054] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the application, but not to limit them; although the application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope defined by the claims of the application.

Claims

1. A source body positioning method based on intelligent gravity and magnetic field separation and Euler deconvolution, characterized in that, The method comprises the following steps: Step 1: According to the geological and physical data of the study area, set the gravity and magnetic anomaly field modeling parameters; Step 2: Randomly generate multiple initial seed points in the three-dimensional grid space of the study area, and expand and grow in a random direction according to the preset rules until Q independent random three-dimensional models of the anomaly field are generated; Step 3: Map the random three-dimensional model position to the real physical coordinate, calculate each model to obtain the forward anomaly field data set; Step 4: Extract the maximum value of the global absolute value of the forward anomaly field data set, and sort all data in ascending order to obtain an ordered model sequence ; Step 5: Extracting the first p data superposition of the ordered model sequence to generate a composite field , the first data as a separate target field , forming a field separation data set ; Step 6: Field separation dataset Train the field separation network until the network reaches a preset convergence criterion; Step 7: Input the collected measured gravity and magnetic exploration data into the trained field separation network, output the separated field data, and calculate the residual field data by subtracting the measured gravity and magnetic exploration data from the separated field data; Step 8: According to the geological interpretation requirements, select the separated field data or the residual field data as the new input, input the trained field separation network for iterative separation to obtain the field separation result; Step 9: According to the field separation result, use the joint Euler deconvolution method to solve and obtain the preliminary source body positioning result; Step 10: Screen the preliminary source body positioning result to finally determine the source body positioning result.

2. The source body positioning method based on intelligent gravity and magnetic field separation and Euler deconvolution according to claim 1, characterized in that, In step 1, based on the geological mapping data and physical property database of the study area, the density and magnetic susceptibility range of different lithology zones are counted, and the physical property parameter interval, initial seed size and maximum growth volume constraint of the anomaly field modeling are set; Define the observation plane position and grid resolution; The anomaly field modeling parameters include gravity gradient data, gravity anomaly data, magnetic anomaly data and magnetic gradient data.

3. The source body positioning method based on intelligent gravity and magnetic field separation and Euler deconvolution according to claim 2, characterized in that, In step 2, the generation of Q independent random three-dimensional models of the anomaly field comprises the following steps: Step 2.1: Establish a growth path stack for each initial seed to record the current growth coordinates; Step 2.2: Determine whether the current growth coordinates of the initial seed meet the growth conditions; Specifically, if the neighborhood grid in the three-dimensional grid space of the initial seed is not occupied and is within the preset growth space, it is marked as a growable area; if the growth is blocked, return to the last growth node and select a growth direction again; Step 2.3: Expand the seed volume in a random direction within the legal growth space until the preset growth number or volume limit is reached; Step 2.4: Loop the above steps until Q independent random three-dimensional models of the anomaly field are generated.

4. The source body positioning method based on intelligent gravity and magnetic field separation and Euler deconvolution according to claim 3, characterized in that, In step 3, the model obtained by growing the random seed in the three-dimensional grid is converted to the real physical space position, and the calculation is as follows: ； In the formula, respectively represent the real physical space position and the grid position corresponding to the i-th grid point in the random three-dimensional model, respectively represent the real physical space position and the grid position corresponding to the i-th grid point in the random three-dimensional model, respectively represent the real physical space position and the grid position corresponding to the i-th grid point in the random three-dimensional model, represents the interval of the real space subdivision grid, represents the minimum value of the grid point coordinate in a certain direction, respectively represent different coordinate axes, represents the grid point coordinate in a certain direction.

5. The source body positioning method based on intelligent gravity and magnetic field separation and Euler deconvolution according to claim 4, characterized in that, In step 3, the gravity or magnetic forward calculation is performed on each random three-dimensional model to obtain the corresponding forward anomaly field data.

6. The source body positioning method based on intelligent gravity and magnetic field separation and Euler deconvolution according to claim 5, characterized in that, The step 5 constitutes the field separation dataset specifically: taking the first forward abnormal field data in the ordered model sequence as network input, the first single separation target field as target output, calculated as follows: ； Composite stacked field And separate target field Forming field separation dataset And divide the training set and the validation set according to the proportion.

7. The source body positioning method based on intelligent gravity and magnetic field separation and Euler deconvolution according to claim 6, characterized in that, The step 6 utilizes the field separation dataset The training of the field separation network is specifically: Construct a field separation network, establish a nonlinear mapping from the composite stacking field to the target separated field through the field separation network, optimize the weight parameters of the field separation network using the divided training set and validation set, and stop until the field separation network reaches the preset convergence standard; The field separation network is composed of a field separation estimation network and a field separation sub-network, the field separation estimation network is composed of multiple convolution layers and activation functions, and is used to extract multi-scale features of the composite stacking field; the field separation sub-network is composed of an encoder for feature extraction and data compression and a decoder for data reconstruction; The calculation formula of the convolution layer of the field separation estimation network is as follows: ； ； ； In the formula, , represents the input data of the current convolutional layer , represents the output result of the current convolutional layer after the convolutional operation on the input data; represents the weight value of the current convolutional layer, represents the number of channels of the input data; , respectively represent the width and depth of the current convolutional layer; and represent the step size of the convolutional operation on the input data and the edge padding value, represents the bias coefficient of the current convolutional layer; represents the size of the convolutional layer , can be represented as: ； In the formula, represents the activation function applied by the current model training.

8. The source body positioning method based on intelligent gravity and magnetic field separation and Euler deconvolution according to claim 7, characterized in that, The step 7 is specifically: inputting the collected measured gravity and magnetic exploration data into the trained field separation network, outputting the separated field data, and calculating the difference between the measured gravity and magnetic exploration data and the separated field data to obtain the residual field data according to the following formula: ； In the formula, is the residual field data, is the collected measured gravity and magnetic exploration data, is the separated field data.

9. The source body positioning method based on intelligent gravity and magnetic field separation and Euler deconvolution according to claim 7, characterized in that, In the step 9, based on the field separation result, a joint Euler deconvolution method is used for solving, which is specifically: A joint Euler deconvolution method for gravity and magnetic anomaly data is proposed by calculating the vertical derivative, and the formula principle is based on the Euler homogeneous equation: ； wherein, with are the observed point and the geologic body center coordinates, respectively, is the structural index, and are the background field and the gravity-magnetic anomaly data, respectively; The vector product form is: ； In order to reduce the selection of the structure index N and the interference of the background field B, the second-order derivative is carried out in the x, y and z directions, and the joint Euler deconvolution formula of the gravity gradient data or the magnetic gradient data is obtained: ； The vector product form is: ； The equation is solved using a sliding window approach, each sliding window solving a corresponding , which is the preliminary source localization result, is the number of solutions.

Images