Geophysical three-dimensional visualization and geologic model construction system

Abstract

The invention discloses a geophysical three-dimensional visualization and geologic model construction system, which belongs to the technical field of geophysical data processing and three-dimensional geologic modeling and comprises a multi-source data acquisition module, a volume rendering module, an interactive segmentation module, an uncertainty expression module, a multi-scale fusion module, an attribute interpolation module and a model construction module. Multi-source geophysical data are converted into a visual three-dimensional visual model through a volume rendering technology, an interactive segmentation algorithm is adopted to accurately extract a geologic body boundary, an uncertainty expression mechanism is innovatively introduced to quantify the reliability of the model, and a multi-scale fusion technology is integrated to realize seamless connection from a regional structure to an ore deposit fine structure. The Kriging interpolation functional product physical property parameter three-dimensional distribution field is provided, the spatial cognitive ability and the decision-making efficiency under the complex geological condition are remarkably improved, and a powerful visual analysis platform is provided for geological exploration and resource evaluation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of geophysical data processing and three-dimensional geological modeling, and particularly relates to a geophysical three-dimensional visualization and geological model construction system based on volume rendering and multi-scale fusion. BACKGROUND

[0002] Geophysical exploration is an important means to understand underground geological structure, evaluate mineral resources and guide engineering construction. With the development of seismic, gravity, electromagnetic and other geophysical exploration technologies, people can obtain massive underground information. However, how to convert these multi-source, heterogeneous geophysical data into intuitive and accurate three-dimensional geological models is still an important challenge in the field of earth science.

[0003] For example, Chinese patent publication CN119273864A discloses a complex geological model generation system based on three-dimensional visualization and GIS technology. The system includes a modeling acquisition module, a terrain labeling module, a boundary extraction module, a boundary verification module and a model generation module. The system obtains the topographic map of the region to be modeled, determines the position of the boundary point, the boundary shape and the terrain features, extracts the boundary features and verifies them, and finally generates a local geological model.

[0004] However, the existing technology has the following shortcomings: first, there is a lack of volume rendering technology for geophysical multi-source data, which cannot intuitively display the spatial distribution characteristics of three-dimensional geophysical fields; second, there is a lack of uncertainty expression mechanism, which fails to quantify the reliability of each part of the model, making it difficult for decision-makers to accurately assess the confidence of the model; third, there is a lack of multi-scale fusion capability, which cannot seamlessly connect from regional tectonic framework to fine structure of ore deposit, limiting the practical value of the model; fourth, there is a lack of attribute interpolation module based on limited observation points, making it difficult to generate continuous three-dimensional distribution of physical properties, affecting the accuracy of resource quantity estimation; fifth, the boundary extraction method relies on simple gradient calculation, which is not accurate enough in complex geological conditions and is prone to misjudgment. SUMMARY

[0005] To solve the above technical problems, the present application provides a geophysical three-dimensional visualization and geological model construction system, which uses volume rendering technology to realize three-dimensional visualization of geophysical data, introduces an uncertainty expression mechanism to quantify model reliability, integrates multi-scale fusion technology to realize cross-scale modeling, provides attribute interpolation function to support resource evaluation, and significantly improves spatial cognition ability and decision-making efficiency under complex geological conditions.

[0006] The technical solution adopted by the present application is: a geophysical three-dimensional visualization and geological model construction system, comprising:

[0007] The multi-source data acquisition module is configured to acquire seismic data, gravity and magnetic data, and electromagnetic data of a region to be modeled, normalize the acquired data to obtain standardized geophysical data if the type of the acquired data matches a preset data type set.

[0008] The volume rendering module is configured to generate a three-dimensional data volume based on the standardized geophysical data, perform volume rendering on the three-dimensional data volume by using a ray casting algorithm to obtain a three-dimensional visualization model with opacity information if the data density in the three-dimensional data volume matches a first density threshold range.

[0009] The interactive segmentation module is configured to receive an interactive instruction of a user for the three-dimensional visualization model, and extract a set of geobody boundary points from a target region by using a region growing algorithm if the boundary definition of the target region indicated by the interactive instruction matches a first definition threshold.

[0010] The uncertainty expression module is configured to calculate a reliability coefficient of each boundary point in the set of geobody boundary points, and express the degree of uncertainty by adjusting the opacity value and color coding value of the corresponding boundary point if the reliability coefficient is lower than a first reliability threshold.

[0011] The multi-scale fusion module is configured to obtain regional tectonic framework data and deposit fine structure data, and perform fusion on data of different scales by using a pyramid layering fusion strategy to generate a cross-scale geological model if the difference between the spatial scale of the regional tectonic framework data and the spatial scale of the deposit fine structure data exceeds a preset scale difference threshold.

[0012] The attribute interpolation module is configured to determine the spatial distribution density of an observation point based on a set of observation point data, and generate a three-dimensional distribution field of a physical property parameter by using a Kriging interpolation method if the spatial distribution density is lower than a first density threshold, and perform local re-interpolation processing on a region with a numerical discontinuity in the generated three-dimensional distribution field if the degree of discontinuity exceeds a preset continuity threshold.

[0013] The model construction module is configured to integrate the three-dimensional visualization model, the set of geobody boundary points, the uncertainty information, the cross-scale geological model, and the three-dimensional distribution field to generate a complete three-dimensional geological model containing reliability annotations.

[0014] In a possible implementation, when acquiring seismic data, the multi-source data acquisition module is configured to perform enhancement processing on the seismic data by using a distributed acoustic sensing technology if the signal-to-noise ratio of the seismic data is lower than a first signal-to-noise ratio threshold, replace traditional geophone data with optical fiber strain rate data, and use the enhanced seismic data as the seismic component of the standardized geophysical data if the signal-to-noise ratio is improved by more than a preset improvement threshold.

[0015] In a possible implementation, when the volume rendering module adopts the ray casting algorithm, a plurality of viewpoint positions are set, a ray is emitted from each viewpoint position to the three-dimensional data volume, if the value of the data voxel through which the ray passes meets a first value range, the color value and the opacity value corresponding to the data voxel are accumulated, if the accumulated opacity reaches a first opacity threshold, the subsequent sampling of the ray is terminated, and a two-dimensional projection image corresponding to the viewpoint is generated.

[0016] In a possible implementation, after receiving the interaction instruction, the interactive segmentation module determines the seed point position indicated by the interaction instruction, takes the seed point as a starting point, if the difference between the geophysical attribute value of the adjacent data point and the attribute value of the seed point is less than a first similarity threshold, the adjacent data point is added to the same geologic body region, and if the number of added data points exceeds a first size threshold, the region growing is stopped and a geologic body boundary point set is output.

[0017] In a possible implementation, when calculating the reliability coefficient, the uncertainty expression module obtains the number and distribution uniformity of the observation points around the boundary point, if the number of observation points is greater than a first number threshold and the distribution uniformity is greater than a first uniformity threshold, the reliability coefficient takes a high value, and if the number of observation points is less than a second number threshold or the distribution uniformity is less than a second uniformity threshold, the reliability coefficient takes a low value.

[0018] In a possible implementation, when adopting the pyramid layering fusion strategy, the multi-scale fusion module takes the region construction framework data as the bottom coarse-scale data and takes the deposit fine structure data as the top fine-scale data, if the ratio of the spatial resolutions of adjacent levels meets a first proportion range, spatial registration between the levels is realized through upsampling or downsampling operation, and if the registration error is less than a first error threshold, weighted fusion of the multi-layer data is performed.

[0019] The present application has the advantages that:

[0020] Firstly, the present application converts geophysical multi-source data into an intuitive three-dimensional visualization model through volume rendering technology, realizes real-time rendering of a three-dimensional data field by adopting a ray casting algorithm, supports interactive exploration, and significantly improves the cognitive efficiency of geologists on underground structures. Compared with the prior art, the present application does not rely on two-dimensional profiles or simple isosurface display, but directly presents the internal structure of the three-dimensional data field by volume rendering, so that the spatial form of the complex geologic body is immediately apparent.

[0021] Secondly, the uncertainty expression mechanism is introduced innovatively, the reliability coefficients of the boundary points are calculated, and the opacity and color coding are used to intuitively show the confidence of each part of the model. This mechanism can help decision makers identify high-uncertainty areas, arrange supplementary exploration work reasonably, avoid making major decisions in low-reliability areas, and reduce exploration risks. The existing technology usually presents the geological model in a deterministic manner, which can easily create the false impression that what you see is what you get, while the uncertainty expression mechanism of the present application can truly reflect the confidence level of the model.

[0022] Thirdly, the present application integrates multi-scale fusion technology and adopts a pyramid hierarchical fusion strategy to realize seamless connection from regional tectonic framework to fine structure of ore deposits. Through spatial registration and weighted fusion, the present application can present macroscopic tectonic pattern and local mineralization details in the same model, meeting the application needs at different levels. Compared with the existing technology which can only establish a single-scale model, the multi-scale fusion capability of the present application greatly improves the application range and practical value of the model.

[0023] Fourthly, the present application provides an attribute interpolation module, which generates a continuous three-dimensional distribution field of physical property parameters based on limited observation point data using the Kriging interpolation method, supporting resource quantity estimation and development plan design. The geological rule constraint is considered in the interpolation process, ensuring the physical rationality of the interpolation results. Compared with the existing technology which lacks attribute interpolation function, the present application can construct a complete physical property parameter model under the condition of sparse observation data, providing a quantitative basis for resource evaluation.

[0024] Fifthly, the present application adopts interactive segmentation technology to accurately extract the boundary of the geological body from the three-dimensional visualization model, and combines the region growing algorithm and user interaction to accurately outline the shape of the geological body under complex geological conditions. Compared with the boundary extraction method of the existing technology which relies on simple gradient calculation, the interactive segmentation technology of the present application has higher precision and stronger adaptability, and can effectively handle the case of fuzzy geological body boundary. BRIEF DESCRIPTION OF DRAWINGS

[0025] Figure 1 is a system architecture diagram of the geophysical three-dimensional visualization and geological model construction system of the present application.

[0026] Figure 2 is a processing flowchart of the multi-source data acquisition module in the system of the present application.

[0027] Figure 3 is a working principle schematic diagram of the volume rendering module in the system of the present application.

[0028] Figure 4 is a calculation logic diagram of the uncertainty expression module in the system of the present application.

[0029] Figure 5is a pyramid hierarchical fusion structure diagram of a multi-scale fusion module in the system of the present application.

[0030] Figure 6 is a Kriging interpolation flow chart of an attribute interpolation module in the system of the present application. DETAILED DESCRIPTION

[0031] Reference will now be made to the drawings, and specific examples relating to the preferred embodiments of the present application will be described in detail. Figures 1-6 The present application will be further described in detail below in conjunction with the accompanying drawings and specific examples. The following examples are used to illustrate the present application, but are not used to limit the scope of the present application.

[0032] Referring to Figure 1 The present application provides a geophysical 3D visualization and geological model construction system, which comprises a multi-source data acquisition module, a volume rendering module, an interactive segmentation module, an uncertainty expression module, a multi-scale fusion module, an attribute interpolation module and a model construction module. The modules are associated through data flow to form a complete geological modeling workflow.

[0033] Referring to Figure 2 The multi-source data acquisition module is responsible for acquiring seismic data, gravity and magnetic data and electromagnetic data of the region to be modeled. In geophysical exploration practice, different physical field methods have different response characteristics to underground media, and comprehensive utilization of multiple methods can obtain more comprehensive geological information.

[0034] Seismic data reflects the wave impedance structure of underground rock layers, and through the propagation time and amplitude characteristics of seismic waves, stratigraphic interfaces and fault structures can be identified. For example, the sampling interval of seismic data is 5 meters, the time sampling rate is 4 milliseconds, and the depth range covered is 0 to 5000 meters.

[0035] Gravity and magnetic data reflect the density and magnetic susceptibility distribution of underground rocks, and are particularly effective for identifying geological bodies with significant density or magnetic differences such as intrusive rocks and mineralized zones. For example, the measurement accuracy of gravity data is 0.01 milligal, and the measurement accuracy of magnetic force data is 0.1 nanotesla.

[0036] Electromagnetic data reflect the resistivity distribution of underground media, and are sensitive to low-resistance bodies such as water-bearing structures, clay layers and mineralized alteration zones. For example, the magnetotelluric sounding method is used, with a frequency range of 320 Hz to 0.0001 Hz, and a detection depth of up to 10000 meters.

[0037] The multi-source data acquisition module first performs type checking on the collected data to confirm whether the data type belongs to the preset data type set. The preset data type set includes seismic reflection data, seismic refraction data, gravity anomaly data, magnetic anomaly data, magnetotelluric data, controlled source audio magnetotelluric data, etc. If the collected data type conforms to the preset data type set, the data is normalized.

[0038] The purpose of normalization processing is to eliminate the influence of different physical dimensions and numerical magnitudes, so that different types of geophysical data can be comprehensively analyzed in a unified numerical space. The normalization formula is as follows:

[0039] ,

[0040] Wherein: is the normalized data value; is the original data value; is the minimum value of the original data; is the maximum value of the original data.

[0041] Through normalization processing, all geophysical data is mapped to the interval of 0 to 1, which is convenient for subsequent data fusion and visualization processing.

[0042] It should be noted that when the signal-to-noise ratio of seismic data is low, it will affect the effect of subsequent processing. The present application introduces distributed acoustic sensing (Distributed Acoustic Sensing, DAS) technology to enhance the processing of seismic data. DAS technology uses optical fiber as a sensing medium, and records seismic wave signals by measuring the strain rate change along the optical fiber. Compared with traditional geophones, DAS technology has the advantages of high spatial resolution, low cost, and is suitable for large-scale deployment.

[0043] For example, when the signal-to-noise ratio of seismic data is lower than the first signal-to-noise ratio threshold (for example, 5 decibels), the system automatically enables the optical fiber strain rate data collected by DAS technology. By integrating and filtering the optical fiber strain rate data, signal quality comparable to or even better than traditional seismic data can be obtained. If the enhanced signal-to-noise ratio improves by more than the preset improvement threshold (for example, 3 decibels), the enhanced seismic data is used as the seismic component of the standardized geophysical data for subsequent volume rendering and modeling processes.

[0044] Referring to Figure 3 , the volume rendering module generates a three-dimensional data volume based on the standardized geophysical data, and uses a ray casting algorithm for volume rendering to obtain a three-dimensional visualization model with opacity information.

[0045] The three-dimensional data volume is a three-dimensional array obtained by regularly gridding the standardized geophysical data according to spatial coordinates. Each grid cell (voxel) corresponds to a small cubic region in space, and stores the geophysical attribute value of the region. For example, the spatial resolution of the three-dimensional data volume is 10 meters x 10 meters x 10 meters, the coverage range is 5000 meters x 5000 meters x 3000 meters, and the total number of voxels contained is 5 x 10^8.

[0046] Volumetric rendering is a visualization technique that directly generates two-dimensional images from a three-dimensional data field without extracting intermediate geometric surfaces, thus preserving the internal structural information of the data field. This invention employs a ray casting algorithm to achieve volumetric rendering.

[0047] The basic principle of the ray casting algorithm is as follows: several rays are emitted from the observer's viewpoint toward the 3D data volume. When each ray passes through the data volume, voxels are sampled along the ray path. The corresponding color and opacity are determined based on the voxel values. Then, the colors and opacities of each voxel are accumulated and synthesized according to the direction of ray propagation to finally obtain the color value of the pixel corresponding to the ray.

[0048] For example, the viewpoint position is set as (x_view, y_view, z_view), and the screen plane is perpendicular to the viewing direction. For each pixel (i, j) on the screen, calculate the ray direction vector from the viewpoint through that pixel pointing into the data body. .

[0049] Along the light With fixed step size (For example Sampling was conducted at a distance of 1 meter, with the sampling points located at:

[0050] ,

[0051] in: The spatial location of the k-th sampling point; Viewpoint position; The sampling point number; This is the sampling step size; This is the direction vector of the light ray.

[0052] For each sampling point Geophysical property values ​​of a point are obtained from a 3D data volume using trilinear interpolation. Then, the attribute values ​​are transferred according to the transfer function. Mapped to color values and opacity value .

[0053] The transfer function is a mapping relationship defined by the user according to geological interpretation needs. For example, for seismic impedance data, the transfer function can be set as follows:

[0054] when (For example When the impedance layer is low, the color is set to blue. Opacity ;

[0055] When ( e.g. ) corresponds to a medium impedance layer, the color is set to green , and the opacity ;

[0056] When , corresponds to a high impedance layer, the color is set to red , and the opacity .

[0057] In the direction of the light ray, the color and opacity are accumulated in a front-to-back compositing manner. The accumulation formula is as follows:

[0058] ,

[0059] ,

[0060] Wherein: is the accumulated color value; is the accumulated opacity; is the color value of the kth sampling point ; is the opacity of the kth sampling point.

[0061] When the accumulated opacity reaches the first opacity threshold ( e.g. ), it means that the light ray has been basically completely absorbed, and the subsequent sampling of the light ray is terminated to improve the rendering efficiency. Finally, the accumulated color value is the color of the screen pixel (i, j).

[0062] By performing the above light ray projection process on all pixels on the screen, a two-dimensional projection image of the three-dimensional data volume, i.e. a three-dimensional visualization model, can be generated. The user can change the viewpoint position and transfer function parameters through interactive operation, update the visualization model in real time, and observe the underground geological structure from different angles and in different display modes.

[0063] It should be noted that before volume rendering, the system will check the data density of the three-dimensional data volume. If the data density is too low (e.g. lower than the first density threshold, set to less than 10 valid data points per 100m x 100m x 100m cube), the data volume is too sparse and it is difficult to generate high-quality rendering images. At this time, the system will prompt the user to supplement the observation data or adjust the rendering parameters.

[0064] The interactive segmentation module is used to extract the boundary of a geological body from a 3D visualization model. A geological body refers to a spatially connected region with similar geophysical properties, such as a certain stratigraphic unit, intrusive rock mass or mineralized zone. Accurate identification of the boundary of a geological body is a key step in building a geological model.

[0065] The present application adopts an interactive segmentation algorithm based on region growing. The user specifies one or more seed points in the 3D visualization model by mouse clicking or other interactive methods, and the seed points are located inside the target geological body. The system expands from the seed point as the starting point, and includes data points with similar geophysical properties in the same geological body region.

[0066] Specifically, the interactive segmentation module first receives the user's interactive instruction to determine the position of the seed point indicated by the interactive instruction . Then, the boundary definition of the region around the seed point is checked. The boundary definition can be measured by calculating the gradient of the geophysical property value around the seed point:

[0067] ,

[0068] Wherein: is the gradient value; is the geophysical property value; , , are the partial derivatives of the attribute value in x, y, z directions respectively.

[0069] If the average gradient value inside the target region is lower than the first definition threshold (for example ), it means that the attribute changes smoothly inside the region, the boundary is clear, and it is suitable for region growing segmentation.

[0070] The steps of the region growing algorithm are as follows:

[0071] 1. Initialization: add the seed point to the processing queue , and mark the seed point as visited;

[0072] 2. Loop processing: take out a data point from the queue , and get its six-neighborhood (up, down, left, right, front, back) data points;

[0073] 3. Similarity judgment: for each neighborhood data point , if it has not been visited, calculate the difference between its attribute value and the attribute value of the seed point : ;

[0074] 4. Conditional expansion: if (the first similarity threshold, e.g. ), indicating that the neighborhood data points belong to the same geobody as the seed point, the neighborhood data points are added to the queue and marked as visited;

[0075] Scale control: the number of marked data points is recorded, and if the number exceeds the first scale threshold (e.g., 10^6 data points), the region growing is stopped to prevent the algorithm from expanding indefinitely;

[0076] 5. Boundary extraction: steps 2 to 5 are repeated until the queue is empty. Finally, all data points marked as visited constitute the target geobody region, and the data points on the region boundary are the geobody boundary point set.

[0077] Through the above interactive segmentation algorithm, the system can automatically identify and extract the geobody boundary according to the user-specified seed point, greatly improving the efficiency and accuracy of boundary delineation.

[0078] Referring to Figure 4 , the uncertainty expression module is used to quantify the reliability of each part of the geological model and visually display the uncertainty distribution through visualization means. Due to limited observation data, approximate interpolation methods, subjective geological interpretation, and other factors, the model has inherent uncertainty. Ignoring uncertainty can increase the risk of decision-making.

[0079] The present application adopts a reliability evaluation method based on data coverage. For each boundary point in the geobody boundary point set, the system calculates its reliability coefficient . The calculation of the reliability coefficient considers two factors: the number of observation points around the boundary point and the spatial distribution uniformity of the observation points.

[0080] The number of observation points reflects the data density around the boundary point. The more observation points there are, the more sufficient the data constraints in this region, and the higher the model reliability. For example, set the first number threshold and the second number threshold .

[0081] The spatial distribution uniformity of the observation points reflects whether the spatial distribution of the observation points is balanced. If the observation points are concentrated in one direction and lack data in other directions, the model has higher uncertainty in the direction lacking data. The uniformity can be measured by calculating the variance of the observation point positions:

[0082] ,

[0083] where: is the uniformity; is the standard deviation of the observation point positions; Maximum possible standard deviation (e.g., the standard deviation when all observation points are concentrated at one point).

[0084] Uniformity The uniformity has a value range of 0 to 1, and the closer to 1 indicates that the distribution is more uniform. For example, a first uniformity threshold is set to 0.9, and a second uniformity threshold is set to 0.8.

[0085] Reliability coefficient The calculation formula of the reliability coefficient is as follows:

[0086] ,

[0087] Wherein: is the reliability coefficient; is the observation point quantity weight; is the uniformity weight; is the observation point quantity normalization function; is the uniformity normalization function.

[0088] Normalization function is defined as follows:

[0089] If , then ;

[0090] If , then ;

[0091] If , then .

[0092] Normalization function is defined as follows:

[0093] If , then ;

[0094] If , then ;

[0095] If , then .

[0096] For example, take , The value range of the reliability coefficient is 0 to 1, and the closer to 1 indicates that the reliability is higher.

[0097] If the reliability coefficient of the boundary point is lower than the first reliability threshold (for example , the boundary point is removed.), which indicates that the point has high uncertainty, and the system expresses the degree of uncertainty by adjusting the opacity value and color coding value of the point.

[0098] Specifically, the opacity value is inversely proportional to the reliability coefficient :

[0099] ,

[0100] wherein: is the uncertainty opacity; is the reliability coefficient.

[0101] The lower the reliability coefficient, the higher the opacity value, and the more transparent the boundary point in the visualization model, which intuitively prompts the user that the model reliability in this area is low.

[0102] In addition, the system can also use color coding to express uncertainty. For example, a cold and warm tone mapping is used:

[0103] High reliability ( ): warm tone (red, orange), indicating sufficient data and reliable model;

[0104] Medium reliability ( ): neutral tone (yellow, green);

[0105] Low reliability ( ): cold tone (blue, purple), indicating sparse data and high model uncertainty.

[0106] Through the joint expression of opacity and color coding, the user can intuitively identify the high uncertainty area in the geological model, providing guidance for supplementing exploration work and avoiding making major decisions in low reliability areas.

[0107] Referring to Figure 5 , the multi-scale fusion module is used to integrate geological data of different spatial scales to achieve seamless connection from regional tectonic framework to fine structure of the deposit. Geological phenomena have obvious scale effect, and regional scale tectonic framework controls the macroscopic distribution of the deposit, while the fine structure inside the deposit determines the grade and shape of the ore body. Single-scale modeling cannot simultaneously consider macroscopic and microscopic information, and multi-scale fusion can build a more comprehensive geological model.

[0108] The present application adopts a pyramid hierarchical fusion strategy. The pyramid model includes multiple levels, and each level corresponds to different spatial resolution. The bottom layer is coarse scale data, which has large coverage but low resolution; the top layer is fine scale data, which has small coverage but high resolution; and the middle level has moderate spatial range and resolution.

[0109] For example, the regional structural framework data is taken as the bottom layer L0, with a spatial resolution of 1000 m x 1000 m and a coverage of 100 km x 100 km; the middle layer L1 has a spatial resolution of 100 m x 100 m and a coverage of 10 km x 10 km; and the top layer L2 is the fine structure data of the deposit, with a spatial resolution of 10 m x 10 m and a coverage of 1 km x 1 km.

[0110] The multi-scale fusion module first checks the spatial scale difference of the data in different layers. The scale difference is defined as the ratio of the spatial resolutions of adjacent layers:

[0111]

[0112] wherein: is the scale ratio; is the spatial resolution of the i-th layer; is the spatial resolution of the i+1-th layer.

[0113] If the scale ratio exceeds a preset scale difference threshold (for example, 2), it indicates that there is a large scale difference between adjacent layers, and spatial registration and fusion are needed.

[0114] Spatial registration is to unify the data in different layers to the same spatial coordinate system. For coarse-scale data, an up-sampling operation is used to improve its resolution to be consistent with that of fine-scale data; for fine-scale data, a down-sampling operation can be used to reduce its resolution to be consistent with that of coarse-scale data, or it can be directly embedded into the corresponding area of coarse-scale data.

[0115] The up-sampling method includes nearest neighbor interpolation, bilinear interpolation, cubic spline interpolation, etc. For example, cubic spline interpolation is used for up-sampling, which can improve the resolution while maintaining the smoothness of the data.

[0116] The registration error is evaluated by calculating the difference of the data in the overlapping area after registration:

[0117]

[0118] wherein: is the registration error; is the number of data points in the overlapping area; is the value of the i-th layer at position ; is the value of the i+1-th layer at position .

[0119] If the registration error is less than a first error threshold (for example, 0.1), it indicates that the registration is successful. ​​​), which indicates that the registration is good and the weighted fusion of multi-layer data can be performed.

[0120] The weighted fusion formula is as follows:

[0121]

[0122] wherein: is the value of the fused data at position ; is the weight coefficient of the i-th layer; is the value of the i-th layer at position ; is the total number of pyramid levels.

[0123] The weight coefficient is determined according to the spatial resolution and data quality of each layer of data. The higher the spatial resolution and the better the data quality of a level, the greater the weight coefficient. For example, the calculation formula of the weight coefficient is as follows:

[0124]

[0125] wherein: is the weight coefficient of the i-th layer; is the spatial resolution of the i-th layer; is the data quality score of the i-th layer; represents the inverse of the resolution, and the higher the resolution, the greater the inverse.

[0126] The data quality score is determined by comprehensively considering the signal-to-noise ratio, coverage density, and measurement accuracy of the data, and the value range is 0 to 1. For example, if the data quality score of a layer is lower than a first quality threshold (for example, 0.5), it indicates that the data quality of the layer is poor, and the system will reduce the weight coefficient of the layer; if the reduced weight coefficient is less than a preset minimum weight value (for example, 0.1), the data of the layer will be removed from the fusion process to avoid the negative impact of low-quality data on the fusion result. Through pyramid layering fusion, the system can simultaneously present the macro-structural pattern and the local fine structure in the same model, generate a cross-scale geological model, and meet the application requirements of different levels.

[0127] Through pyramid layering fusion, the system can simultaneously present the macro-structural pattern and the local fine structure in the same model, generate a cross-scale geological model, and meet the application requirements of different levels.

[0128] For example, see Figure 6 ​​​The attribute interpolation module is used for generating continuous three-dimensional distribution field of the physical property parameters based on limited observation point data, and supports resource quantity estimation and development scheme design. In geological exploration practice, the data obtained by direct observation means such as drilling and tunneling are often discrete and limited, while resource evaluation and engineering design need to understand the physical property parameter distribution of the entire region. The attribute interpolation can infer the parameter values of unobserved positions according to the observation point data, and construct a continuous three-dimensional distribution model.

[0129] The present application adopts Kriging interpolation method. Kriging interpolation is an optimal linear unbiased estimation method based on geostatistics, which can estimate the parameter values of unobserved positions on the basis of considering spatial correlation, and give the estimation error.

[0130] The attribute interpolation module first acquires an observation point data set, including the spatial coordinates of each observation point and the physical property parameter value . Then, the spatial distribution density of the observation points is calculated:

[0131] ,

[0132] Among them: is the spatial distribution density; is the number of observation points; is the volume of the region to be modeled.

[0133] If the spatial distribution density is lower than the first density threshold (for example per cubic meter), it indicates that the observation points are relatively sparse, and the spatial correlation information needs to be fully utilized in the interpolation process.

[0134] The core of Kriging interpolation is the semivariogram model. The semivariogram describes the variability of the physical property parameters in space and the relationship with the distance:

[0135] ,

[0136] Among them: is the semivariogram value when the distance is ; is the number of observation point pairs with a distance of ; is the physical property parameter value of position ; is the physical property parameter value of the position .

[0137] ​According to the empirical semi-variogram, a theoretical semi-variogram model is fitted. Commonly used theoretical models include spherical model, exponential model, Gaussian model, etc. For example, the spherical model is used:

[0138] ,

[0139] wherein: is the semi-variogram value; is the nugget effect; is the sill; is the range; is the distance.

[0140] The nugget effect reflects the small-scale variation and measurement error; the sill reflects the overall variation degree; the range reflects the action distance of spatial correlation, and the parameter value is no longer related when the distance exceeds the range.

[0141] The least squares method is used to fit the theoretical semi-variogram model, and the optimal estimation values of the parameters , and are obtained. The goodness of fit is used to evaluate the fitting effect:

[0142] ,

[0143] wherein: is the goodness of fit; is the distance order; is the observed semi-variogram value; is the semi-variogram value of the theoretical model; is the average value of the observed semi-variogram.

[0144] If the goodness of fit is greater than the first fitting threshold (for example, ), it indicates that the theoretical model can well describe the spatial variability of the data, and the Kriging interpolation can be performed based on the model.

[0145] For the position to be interpolated, the Kriging interpolation estimated value is a linear combination of the observed point values:

[0146] ,

[0147] wherein: is the interpolation estimated value of the position ; and is the weight coefficient of the i-th observation point. is the value of the property parameter of the i th observation point; is the total number of observation points.

[0148] weight coefficient By solving the Kriging equation set:

[0149] ,

[0150] ,

[0151] wherein: is the semi-variation function value between the observation point and the observation point ; is the distance between the observation point and the observation point ; is the Lagrange multiplier; is the semi-variation function value between the observation point and the point to be interpolated; is the distance between the observation point and the point to be interpolated.

[0152] Solving the above equation set, the weight coefficient is obtained, and then the interpolation estimation value is calculated.

[0153] Kriging interpolation can also give the estimation error (Kriging variance):

[0154] ,

[0155] wherein: is the Kriging variance; is the weight coefficient; is the semi-variation function value between the observation point and the point to be interpolated; is the Lagrange multiplier.

[0156] The Kriging variance reflects the uncertainty of the interpolation estimation, and the smaller the variance, the more reliable the estimation.

[0157] Perform Kriging interpolation on all grid points in the region to be modeled, and a three-dimensional distribution field of the property parameter can be generated.

[0158] It should be noted that the interpolation result needs to meet the physical rationality. If the property parameter value obtained by interpolation exceeds the physically reasonable range (for example, the density is less than 0 or the porosity is greater than 1), it indicates that there is a problem in the interpolation process. At this time, the system will perform constrained interpolation processing on the point, introduce upper and lower bound constraints, and ensure that the interpolation result conforms to the geological law.

[0159] In addition, if there is a numerical discontinuity region (e.g., the parameter values of adjacent grid points differ by more than a preset continuity threshold, such as ), it indicates that there may be a geological fault or data missing in this region. The system will perform local re-interpolation processing on this region, using higher sampling density or introducing geological constraints to improve interpolation accuracy.

[0160] The model construction module is responsible for integrating the outputs of the aforementioned modules to generate a complete three-dimensional geological model with reliability annotations. This module unifies the management of three-dimensional visualization models, geological body boundary point sets, uncertainty information, cross-scale geological models, and three-dimensional distribution fields to form a comprehensive geological model database.

[0161] The model construction module first establishes a unified spatial coordinate system to ensure that all data is in the same coordinate frame. Then, the area to be modeled is divided into regular three-dimensional grids, with the grid resolution determined according to application requirements, for example, 10m x 10m x 10m.

[0162] For each grid cell (spatial unit), the model construction module assigns an attribute vector containing the following information:

[0163] Geophysical properties: from the three-dimensional visualization model, including seismic wave velocity, density, resistivity, etc.

[0164] Geological properties: from the geological body boundary point set, identifying the geological body type (stratum unit, rock mass, ore body, etc.) to which the unit belongs;

[0165] Uncertainty indicators: from the uncertainty expression module, quantifying the reliability of the model for this unit;

[0166] Data source identification: records the main source of data for this unit (direct observation, interpolation calculation, cross-scale fusion, etc.).

[0167] The model construction module also supports visualization of the model. When visualizing, if the uncertainty indicator of a spatial unit is higher than a second reliability threshold (e.g., the uncertainty indicator ), the system will add special markers (e.g., using grid wireframes, flashing effects, or color warnings) to this unit, prompting users that the model reliability in this area is low and needs to be used with caution.

[0168] In addition, the model construction module provides model export functions, supporting the export of geological models to various common formats such as VTK (Visualization Toolkit) format, GOCAD format, Petrel format, etc., facilitating interaction with other geological modeling software and numerical simulation software.

[0169] In one possible implementation, the system also includes a surface reconstruction module for extracting smooth boundary surfaces from the set of boundary points of the geological body. Boundary surfaces are the outer surfaces of the geological body in three-dimensional space and are of great significance for geological body volume calculation, resource estimation, and visualization.

[0170] The surface reconstruction module first checks the spatial sparsity of the boundary points. Sparsity is defined as the ratio of the average spacing between boundary points to the size of the geological body features:

[0171] ,

[0172] in: Sparsity; The average spacing between boundary points; The characteristic dimensions of a geological body (e.g., the maximum length or thickness of the geological body).

[0173] If sparsity Exceeding the first sparsity threshold (e.g.) This indicates that the boundary points are sparsely distributed, and directly connecting the boundary points may result in an uneven surface. In this case, the system uses the Delaunay triangulation algorithm to generate a triangular mesh surface.

[0174] Delaunay triangulation is a method for connecting discrete point sets into triangular meshes. It maximizes the minimum angle and can generate triangles with good shapes. For a set of boundary points in 3D space, the system first projects the boundary points onto a local tangent plane, performs Delaunay triangulation on the 2D plane, and then maps the triangle vertices back to 3D space to form a triangular mesh surface.

[0175] The generated triangular mesh needs to meet certain geometric quality standards. The system checks the minimum angle of each triangle. ,like Less than a preset angle threshold (e.g.) This indicates that the triangle is too flat, which affects the smoothness of the surface and the accuracy of calculations. The system optimizes and adjusts these low-quality triangles by using mesh optimization techniques such as edge flipping and vertex insertion to improve the quality of the triangles.

[0176] The optimized triangular mesh surface has good smoothness and geometric accuracy, and can accurately represent the spatial morphology of geological bodies.

[0177] The implementation process of the present invention will be further illustrated below through a specific embodiment.

[0178] A certain mining area needs to establish a three-dimensional geological model to guide the development of the mine. The area of the mining area is 5 square kilometers, and the exploration depth is 2000 meters. The existing data includes:

[0179] Seismic reflection data: covering the whole area, spatial sampling interval 10 meters, time sampling rate 4 milliseconds;

[0180] Gravity data: 200 measurement points, average interval 100 meters;

[0181] Magnetic data: 150 measurement points, average interval 120 meters;

[0182] Drilling data: 30 drill holes, obtaining physical parameters such as lithology, density, and porosity.

[0183] First, the multi-source data acquisition module normalizes the seismic data and gravity and magnetic data. The signal-to-noise ratio of the seismic data is evaluated to be 4.5 decibels, which is lower than the first signal-to-noise ratio threshold of 5 decibels. The system uses DAS technology to collect optical fiber strain rate data for enhancement processing. After enhancement, the signal-to-noise ratio is improved to 7.8 decibels, with an improvement of 3.3 decibels, which exceeds the preset improvement threshold of 3 decibels. The enhanced seismic data is used as the seismic component of the standardized geophysical data.

[0184] Then, the volume rendering module generates a three-dimensional data volume based on the standardized geophysical data, with a spatial resolution of 10 meters x 10 meters x 10 meters. The volume rendering is performed using the ray casting algorithm, with the viewpoint set to 2000 meters above the center of the mining area. The transfer function maps the seismic wave impedance to color and opacity. The three-dimensional visualization model generated by rendering clearly shows the spatial distribution of underground strata and fault structures.

[0185] Geologists specify seed points in the visualization model to designate ore body regions through the interactive segmentation module. The system uses a region growing algorithm to automatically extract the boundaries of the ore bodies, identifying a total of 3 main ore bodies, with the boundary point set containing about 50,000 data points.

[0186] The uncertainty expression module calculates the reliability coefficient of each boundary point. The number of observation points around ore body 1 is 15, with a uniformity of 0.75, and the reliability coefficient is calculated as , which is high; the number of observation points around ore body 2 is 5, with a uniformity of 0.45, and the reliability coefficient is calculated as , which is lower than the first reliability threshold of 0.5. The system sets the opacity of the ore body 2 boundary to 0.56 and the color coding to a cool tone, which directly indicates that the uncertainty of this ore body is high.

[0187] The multi-scale fusion module integrates regional structure framework data (spatial resolution 1000 meters) and deposit fine structure data (spatial resolution 10 meters), and the scale ratio is 100, which exceeds the preset scale difference threshold 5. The system adopts a pyramid hierarchical fusion strategy, establishes three levels of pyramid models, generates a cross-scale geological model through upsampling and weighted fusion, and can not only show the macro pattern of regional fault structure controlling the distribution of deposits, but also present the fine structure of the internal grade change of the ore body.

[0188] The attribute interpolation module generates a three-dimensional distribution field of density based on the physical property parameter data of 30 drill holes by using the Kriging interpolation method. The fitting degree of the semi-variogram function is greater than the first fitting degree threshold 0.8. The density distribution field generated by interpolation has a spatial resolution of 10 meters x 10 meters x 10 meters and covers the whole area. During the interpolation process, the system found that the density difference between adjacent grid points in a certain area reached 0.8 g / cm3, which exceeded the preset continuity threshold 0.5. Local re-interpolation processing was performed on this area, and geological fault constraints were introduced to improve the interpolation accuracy.

[0189] Finally, the model construction module integrates all data to generate a complete three-dimensional geological model containing geophysical attributes, geological attributes, uncertainty indicators and data source identifiers. The model contains 5x10^7 spatial units, and the attribute vector length of each unit is 20. The model is exported in VTK format for subsequent resource estimation and mining design.

[0190] The present embodiment successfully integrates multi-source geophysical data and drilling data by using the geophysical three-dimensional visualization and geological model construction system, and constructs a high-precision, cross-scale and uncertainty-annotated three-dimensional geological model, which provides a scientific basis for mine development and significantly improves the accuracy of resource evaluation and the reliability of decision-making.

[0191] The above only describes the preferred embodiments of the present application and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A geophysical 3D visualization and geological model building system, characterized in that, The system comprises: a multi-source data acquisition module for acquiring seismic data, gravity and magnetic data and electromagnetic data of a region to be modeled, and performing normalization processing on the data to obtain standardized geophysical data if the type of the acquired data conforms to a preset data type set; a volume rendering module for generating a three-dimensional data volume based on the standardized geophysical data, and performing volume rendering on the three-dimensional data volume using a ray casting algorithm to obtain a three-dimensional visualization model with opacity information if the data density in the three-dimensional data volume conforms to a first density threshold range; an interactive segmentation module for receiving an interactive instruction of a user for the three-dimensional visualization model, and extracting a set of geological body boundary points from a target region using a region growing algorithm if the boundary definition of the target region indicated by the interactive instruction conforms to a first definition threshold; an uncertainty expression module for calculating a reliability coefficient of each boundary point in the set of geological body boundary points, and expressing the degree of uncertainty by adjusting the opacity value and color coding value of the corresponding boundary point if the reliability coefficient is lower than a first reliability threshold; a multi-scale fusion module for obtaining regional tectonic framework data and deposit fine structure data, and performing fusion on data of different scales using a pyramid layering fusion strategy if the difference between the spatial scale of the regional tectonic framework data and the spatial scale of the deposit fine structure data exceeds a preset scale difference threshold to generate a cross-scale geological model; an attribute interpolation module for determining the spatial distribution density of an observation point based on a set of observation point data, and generating a three-dimensional distribution field of physical property parameters using a Kriging interpolation method if the spatial distribution density is lower than a first density threshold, and performing local re-interpolation processing on a region with a numerical discontinuity in the generated three-dimensional distribution field if the degree of discontinuity exceeds a preset continuity threshold; a model construction module for integrating the three-dimensional visualization model, the set of geological body boundary points, the uncertainty information, the cross-scale geological model and the three-dimensional distribution field to generate a complete three-dimensional geological model containing reliability annotations.

2. The geophysical three-dimensional visualization and geological model construction system according to claim 1, wherein: the multi-source data acquisition module, when acquiring seismic data, performs enhancement processing on the seismic data using a distributed acoustic sensing technology if the signal-to-noise ratio of the seismic data is lower than a first signal-to-noise ratio threshold, replaces traditional geophone data with optical fiber strain rate data, and uses the enhanced seismic data as the seismic component of the standardized geophysical data if the signal-to-noise ratio improvement amplitude exceeds a preset improvement threshold.

3. The geophysical three-dimensional visualization and geological model construction system according to claim 1, wherein: the volume rendering module, when using the ray casting algorithm, sets a plurality of viewpoint positions, emits light rays from each viewpoint position to the three-dimensional data volume, accumulates the color value and opacity value of the data voxel through which the light ray passes if the value of the data voxel conforms to a first value range, and terminates subsequent sampling of the light ray if the accumulated opacity reaches a first opacity threshold to generate a two-dimensional projection image corresponding to the viewpoint.

4. The geophysical three-dimensional visualization and geological model construction system according to claim 1, wherein: The interactive segmentation module determines the seed point position indicated by the interaction instruction after receiving the interaction instruction, takes the seed point as a starting point, and adds adjacent data points to the same geological body region if the difference between the geophysical attribute value of the adjacent data points and the attribute value of the seed point is less than a first similarity threshold. If the number of added data points exceeds a first size threshold, the region growing is stopped and a set of geological body boundary points is output.

5. The geophysical three-dimensional visualization and geological model construction system of claim 1, wherein: The uncertainty expression module obtains the number and distribution uniformity of observation points around the boundary point when calculating the reliability coefficient. If the number of observation points is greater than a first number threshold and the distribution uniformity is greater than a first uniformity threshold, the reliability coefficient takes a high value. If the number of observation points is less than a second number threshold or the distribution uniformity is less than a second uniformity threshold, the reliability coefficient takes a low value.

6. The geophysical three-dimensional visualization and geological model construction system of claim 1, wherein: The multi-scale fusion module takes the regional tectonic framework data as the bottom layer coarse scale data and the deposit fine structure data as the top layer fine scale data when adopting the pyramid layering fusion strategy. If the ratio of the spatial resolution of adjacent levels meets a first proportion range, spatial registration between levels is realized through upsampling or downsampling operations. If the registration error is less than a first error threshold, weighted fusion of multi-layer data is performed.

7. The geophysical three-dimensional visualization and geological model construction system of claim 1 or 6, wherein: The multi-scale fusion module determines the weight coefficient according to the spatial resolution and data quality of each layer of data when performing weighted fusion of multi-layer data. If the data quality score of a certain layer is lower than a first quality threshold, the weight coefficient of that layer is reduced. If the reduced weight coefficient is less than a preset minimum weight value, the data of that layer is removed from the fusion process.

8. The geophysical three-dimensional visualization and geological model construction system of claim 1, wherein: The attribute interpolation module establishes a semi-variogram function model when adopting the Kriging interpolation method. If the fitting goodness of the semi-variogram function is greater than a first fitting degree threshold, the physical property parameter value of the point to be interpolated is calculated based on the semi-variogram function model. If the calculated physical property parameter value exceeds the physically reasonable range, the point is subjected to constrained interpolation processing to ensure that the interpolation result conforms to the geological regularity.

9. The geophysical three-dimensional visualization and geological model construction system of claim 1, wherein: The model construction module assigns an attribute vector containing geophysical attributes, geological attributes, uncertainty indicators, and data source identifiers to each spatial cell in the model when generating a complete three-dimensional geological model. If the uncertainty indicator of a certain spatial cell is higher than a second reliability threshold, a special marker is added to the cell during model visualization, prompting the user that the model reliability of this region is low.

10. The geophysical three-dimensional visualization and geological model construction system of claim 1, wherein: The system further comprises a surface reconstruction module, configured to extract a boundary surface from the set of boundary points of the geological body, if the spatial distribution of the boundary points is sparse to a degree exceeding a first sparsity threshold, generate a triangular mesh surface using a Delaunay triangulation algorithm, and if the smallest angle of the triangular mesh is smaller than a preset angle threshold, optimize and adjust the mesh to ensure the smoothness and geometric accuracy of the surface.

Images