Cloud-computing-based deep geothermal resource exploration data processing platform

Abstract

This invention relates to the field of exploration data processing technology, specifically to a cloud-based deep geothermal resource exploration data processing platform. This platform includes a first acquisition module for acquiring spatial data points and their corresponding sets of physical parameters; a second acquisition module for acquiring core data points; and a clustering module for clustering all spatial data points output from the inversion of the target exploration area based on all core data points. This invention improves the reliability of spatial data point clustering, thereby enabling the allocation of data points from the edge areas of key regions such as reservoirs, fault zones, and hydrothermal channels into clusters related to these key regions. This, in turn, improves the accuracy of subsequent analysis and identification of geothermal resources in the explored area.

Description

Technical Field

[0001] This invention relates to the field of exploration data processing technology, specifically to a cloud computing-based deep geothermal resource exploration data processing platform. Background Technology

[0002] Deep geothermal resource exploration refers to the process of detecting, assessing, and analyzing the development potential of geothermal resources existing at deeper levels of the Earth's crust (usually 2,000 to 10,000 meters or even deeper). Currently, in the field of deep geothermal resource exploration, an initial model, such as a three-dimensional resistivity model, is usually constructed based on data such as the geological scale and structural complexity of the surveyed area. After the initial model is constructed, the MT inversion algorithm (Occam inversion) is used to perform inversion, and the traditional DBSCAN clustering algorithm is used to analyze the output inversion results. Based on clustering analysis, not only can spatially anomalous distribution structures be extracted, such as key geological information such as underground fault zones, hydrothermal channels, or geothermal reservoirs, but it can also provide more intuitive and quantitative analytical basis for subsequent delineation of geothermal reservoir candidate areas, selection of target areas, and exploration deployment.

[0003] Traditional DBSCAN clustering algorithms primarily determine core points based on the number of data points within a neighborhood, and then perform clustering based on these core points. In other words, traditional DBSCAN clustering typically groups data points with similar physical (electrical) properties into one cluster, and uses the clustering results to extract geological information such as underground fault zones, hydrothermal channels, or geothermal reservoirs. However, this method of determining core points can lead to incomplete or unreliable clustering results. Unreliable clustering results can result in low reliability or accuracy in subsequent analyses of underground geological structures and resource distribution, such as geothermal reservoirs. Marginal regions are often affected by discontinuities in adjacent rock strata or structures, leading to sudden changes in the physical properties of geothermal reservoirs. These changes can cause data points from these marginal regions to fail to be assigned to clusters associated with the geothermal reservoir when using traditional clustering methods. This inability to assign data points from the marginal regions to clusters related to the geothermal reservoir can result in errors in subsequent analysis or assessment of the underground geothermal reservoir distribution. Therefore, improving the clustering accuracy of the inversion output data points after inversion to ensure the accuracy of subsequent analysis or identification of geothermal resources in the surveyed area is a pressing issue that needs to be addressed. Summary of the Invention

[0004] To address the aforementioned problems, this invention provides a cloud computing-based deep geothermal resource exploration data processing platform, the specific technical solution of which is as follows:

[0005] One embodiment of the present invention provides a cloud computing-based deep geothermal resource exploration data processing platform, which includes a power grid business data management service terminal comprising:

[0006] The first acquisition module is used to acquire all spatial data points output by the inversion of the target survey area and the set of physical property parameters corresponding to the spatial data points.

[0007] The second acquisition module is used to obtain a first data point and a data point to be judged from the spatial data points based on the minimum number within a preset neighborhood and a preset neighborhood radius; to obtain the physical property parameter feature value of the data point to be judged based on the set of physical property parameters of all spatial data points within the preset neighborhood radius of the data point to be judged; to obtain the structural distribution feature value of the data point to be judged based on the distribution characteristics of the spatial data points within the preset neighborhood radius of the data point to be judged; to obtain the comprehensive density attribute feature value corresponding to the data point to be judged based on the physical property parameter feature value and the structural distribution feature value; and to obtain a second data point from the data points to be judged based on the comprehensive density attribute feature value corresponding to the data point to be judged, wherein both the first data point and the second data point belong to core data points.

[0008] The clustering module is used to cluster all spatial data points output from the inversion of the target survey area based on all core data points among all spatial data points.

[0009] Beneficial effects: This invention includes a first acquisition module, used to acquire all spatial data points output from the inversion of the target survey area and the set of physical property parameters corresponding to the spatial data points; a second acquisition module, used to obtain a first data point and a data point to be judged from the spatial data points according to the minimum number within a preset neighborhood and a preset neighborhood radius, to obtain the physical property parameter feature value of the data point to be judged according to the set of physical property parameters of all spatial data points within the preset neighborhood radius of the data point to be judged, to obtain the structural distribution feature value of the data point to be judged according to the distribution characteristics of the spatial data points within the preset neighborhood radius of the data point to be judged, to obtain the comprehensive density attribute feature value corresponding to the data point to be judged according to the physical property parameter feature value and the structural distribution feature value, and to obtain a second data point from the data points to be judged according to the comprehensive density attribute feature value corresponding to the data point to be judged, wherein both the first data point and the second data point belong to core data points; and a clustering module, used to cluster all spatial data points output from the inversion of the target survey area according to all core data points among all spatial data points. Furthermore, based on the comprehensive density attribute feature value corresponding to the data point to be judged, this invention can improve the accuracy of dividing data points in the edge areas of key regions such as geothermal reservoirs, fault zones, and hydrothermal channels. Alternatively, based on the comprehensive density attribute feature value corresponding to the data point to be judged, this invention can divide data points in the edge areas of key regions such as geothermal reservoirs, fault zones, and hydrothermal channels into clusters related to key regions such as geothermal reservoirs, fault zones, and hydrothermal channels as much as possible. This can improve the accuracy of subsequent analysis or identification of geothermal resources in the surveyed area. In other words, based on the comprehensive density attribute feature value corresponding to the data point to be judged, this invention can improve the accuracy of clustering. Attached Figure Description

[0010] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a structural block diagram of a cloud computing-based deep geothermal resource exploration data processing platform according to the present invention. Detailed Implementation

[0012] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention are within the protection scope of the embodiments of the present invention.

[0013] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art.

[0014] This embodiment provides a cloud computing-based deep geothermal resource exploration data processing platform, which is described in detail below:

[0015] like Figure 1 As shown in the figure, this embodiment provides a cloud computing-based deep geothermal resource exploration data processing platform, including:

[0016] The first acquisition module 01 is used to acquire all spatial data points output by the inversion of the target survey area and the set of physical property parameters corresponding to the spatial data points.

[0017] The main objective of this embodiment is to improve the accuracy of subsequent analysis or identification of geothermal resources in the surveyed area by enhancing the accuracy of clustering. This embodiment achieves this improvement by altering the method for determining the core data points; specifically, compared to the traditional DBSCAN clustering process, this embodiment only changes the process for determining the core data points and does not alter other steps in the clustering process. Furthermore, for ease of understanding, this embodiment will subsequently describe the clustering process of spatial data points output after inversion in any deep geothermal resource exploration area as an example, and this deep geothermal resource exploration area will be designated as the target exploration area.

[0018] This embodiment first uses the MT inversion algorithm to invert the target survey area, and obtains all spatial data points output from the inversion of the target survey area, as well as the set of physical property parameters corresponding to the spatial data points. The output spatial data points are three-dimensional data points. The coordinates of the first dimension of any spatial data point are the planar coordinates of the longitude of the corresponding spatial data point after being transformed by geographic coordinate projection; the coordinates of the second dimension are the planar coordinates of the dimension of the corresponding spatial data point after being transformed by geographic coordinate projection; and the coordinates of the third dimension are the depth of the corresponding spatial data point. Moreover, the coordinates of the first and second dimensions of the spatial data point can also be the depth of the corresponding spatial data point. The longitude and latitude of the UTM coordinates are projected. The unit of each coordinate value in the spatial data point is consistent, that is, the unit of each coordinate value in the spatial data point is meters. The coordinates of the output spatial data points usually represent the spatial distribution of different locations in the survey area. The parameters in the set of physical property parameters corresponding to the output spatial data points include, but are not limited to, resistivity, resistivity gradient, etc. Moreover, the types and number of physical property parameters that make up the set of physical property parameters corresponding to different spatial data points are consistent. That is, if the set of physical property parameters corresponding to a certain spatial data point is composed of resistivity and resistivity gradient, then all the sets of physical property parameters corresponding to all spatial data points in this embodiment have resistivity and resistivity gradient, and the parameter types at the same location in different sets of physical property parameters are also the same.

[0019] In addition, magnetotellurics (MT) is the most common physical exploration method. It typically measures changes in the natural electromagnetic field at the Earth's surface, especially the relative changes in electric and magnetic fields at different frequencies, to invert the distribution of subsurface electrical or physical parameters. This helps reveal information such as subsurface geological structure and resource distribution. In other words, inversion using inversion algorithms to invert the physical parameters or spatial structure of the subsurface medium is a quantitative technique process that uses observational data to deduce the physical parameters or spatial structure of the subsurface medium. The inversion process using inversion algorithms also includes data acquisition and preprocessing, initial model construction and constraints, etc. The multi-source data acquired during data acquisition includes seismic waveform data (amplitude, phase, travel time, etc.) of the target survey area. The data includes well logging data (sonic transit time, density, etc.), geological prior information (stratigraphic columnar section, structural interpretation results), etc. Data preprocessing mainly involves standardizing the collected data. The initial model is based on regional geological patterns (such as sedimentary patterns and fault distribution) and well logging data. Model constraints refer to the introduction of 3D seismic structural interpretation results as spatial morphological constraints. Each grid cell in the initial model is an inversion cell with a clear spatial location and its corresponding electrical or physical parameters (such as resistivity). Since the process of using inversion algorithms to invert the survey area and obtain inversion results is a well-known technology, the specific process of obtaining inversion results will not be described in this embodiment.

[0020] Therefore, this embodiment can obtain all spatial data points output by the inversion of the target survey area and the set of physical property parameters corresponding to the spatial data points through the above process.

[0021] The second acquisition module 02 is used to obtain a first data point and a data point to be judged from the spatial data points based on the minimum number within a preset neighborhood and a preset neighborhood radius; to obtain the physical property parameter feature value of the data point to be judged based on the set of physical property parameters of all spatial data points within the preset neighborhood radius of the data point to be judged; to obtain the structural distribution feature value of the data point to be judged based on the distribution characteristics of the spatial data points within the preset neighborhood radius of the data point to be judged; to obtain the comprehensive density attribute feature value corresponding to the data point to be judged based on the physical property parameter feature value and the structural distribution feature value; and to obtain a second data point from the data points to be judged based on the comprehensive density attribute feature value corresponding to the data point to be judged. Both the first data point and the second data point belong to core data points.

[0022] Currently, when using the traditional DBSCAN clustering algorithm to cluster all spatial data points output from the inversion of a target survey area, it primarily relies on the number of data points within the neighborhood to determine core points, and then performs clustering based on these core points. In other words, the traditional DBSCAN clustering algorithm typically groups data points with similar physical properties (i.e., electrical parameters) into one class, and uses the clustering results to extract geological information such as underground fault zones, hydrothermal channels, or geothermal reservoirs. However, this method of determining core points can lead to incomplete or unreliable clustering results. The absence of reliable data can lead to low reliability or accuracy in subsequent analyses or identifications of underground geothermal resource distribution. For example, the edges of critical areas such as reservoirs, fault zones, and hydrothermal conduits are often affected by discontinuities in adjacent rock strata or structures, causing sudden changes in the physical properties of these edges. This can result in data points in these edges not being assigned to clusters related to the critical areas when using traditional clustering methods. Consequently, the completeness and reliability of the obtained clusters are low. The marginal areas of key regions often exhibit structural features such as tectonic fracture zones and stratigraphic faulting due to the presence of numerous lithological transition zones. These features result in a scattered distribution of data points in these marginal areas. This distribution characteristic can lead to situations where, when using traditional clustering methods, data points in these marginal areas cannot be assigned to clusters related to the key regions. Consequently, the completeness and reliability of the obtained clusters are low. Low completeness and reliability of the clustering results can lead to lower accuracy in subsequent analysis or identification of geothermal resources in the surveyed area. If data points in the edge region of a geothermal reservoir cannot be assigned to clusters related to the reservoir region, the subsequent identification of geothermal reservoir regions based on clustering results will suffer from low completeness. To improve the accuracy and completeness of clustering and ensure the reliability and accuracy of subsequent analysis or identification of underground geothermal resource distribution, this embodiment will determine the core data points among all spatial data points output from the inversion of the survey area based on the minimum number of data points in the neighborhood, the physical property characteristic values ​​of the data points, and the structural distribution characteristic values. In other words, this embodiment will improve the accuracy of clustering by changing the method of determining the core data points.

[0023] Therefore, based on the above analysis, it can be seen that the main purpose of this embodiment is to obtain the core data points in the spatial data points. In this embodiment, the process of obtaining the core data points is as follows: first, based on the minimum number in the preset neighborhood and the preset neighborhood radius, the first data point and the data point to be judged in the spatial data points are obtained. Then, the data point to be judged is re-judged as a core data point to obtain the second data point in the data point to be judged. Both the first data point and the second data point belong to the core data points, that is, all the core data points in the spatial data points are obtained.

[0024] Based on the core data point acquisition process described above, this embodiment first needs to obtain the first data point and the data point to be judged in the spatial data points. The specific acquisition process for the first data point and the data point to be judged in the spatial data points is as follows:

[0025] First, the DBSCAN parameters are obtained. In other words, this embodiment first needs to obtain the preset neighborhood radius and the minimum number of points within the preset neighborhood. In this embodiment, the preset neighborhood radius and the minimum number of points within the preset neighborhood need to be set according to the actual situation. For example, since the spacing between conventional stations in MT exploration is usually between 200 and 500 meters, in order to ensure that the neighborhood of each point can cover the adjacent points sufficiently without excessive diffusion to a large area, which is conducive to the discovery of local boundary structures, this embodiment can set the preset neighborhood radius to 100 meters. Since the grid point spacing is usually around tens of meters after MT inversion, and the number of points within a 100-meter radius is generally between 20 and 30, and since these areas are usually dense for stable structural areas within geothermal bodies, in order to ensure that only areas with sufficiently dense data points in local space are marked as the first data point, thereby forming an effective cluster, this embodiment can set the minimum number of points within the preset neighborhood to 10.

[0026] The specific process of obtaining the first data point and the data point to be judged in the spatial data points based on the minimum number and the preset neighborhood radius is as follows: For any spatial data point, if the number of spatial data points within the preset neighborhood radius of the spatial data point is less than the minimum number within the preset neighborhood, then the spatial data point is recorded as the data point to be judged; otherwise, the spatial data point is recorded as the first data point. That is, the process of determining the first data point and the data point to be judged based on the minimum number and the preset neighborhood radius is consistent with the existing process of determining core data points and noisy data points or isolated data points based on the minimum number and the preset neighborhood radius. The first data point belongs to the core data point.

[0027] Since the data points to be judged may include data points corresponding to the edge regions of key areas such as thermal reservoirs, fault zones, and hydrothermal channels, this embodiment re-judges the data points after obtaining them. Specifically, after obtaining the data points, this embodiment obtains the physical property parameter characteristic values ​​of the data points based on the set of physical property parameters of all spatial data points within a preset neighborhood radius. These physical property parameter characteristic values ​​are one of the key parameters for subsequently obtaining the core data points among the data points to be judged. The specific process for obtaining the physical property parameter characteristic values ​​of the data points to be judged is as follows:

[0028] For any data point A to be judged: First, obtain all spatial data points within a preset neighborhood radius of data point A. Denote the set of all spatial data points within this radius as the neighborhood set of data point A. Then, denote all spatial data points within this neighborhood set as neighborhood data points of data point A. The neighborhood set of data point A includes data point A itself. Next, based on the set of physical property parameters of the neighborhood data points of data point A, obtain the neighborhood physical property parameter difference value corresponding to each physical property parameter in the set of physical property parameters of data point A. Finally, obtain the normalized sum of the neighborhood physical property parameter difference values ​​corresponding to all physical property parameters in the set of physical property parameters of data point A, and use this sum as the value of the data point to be judged. The first index value of point A is obtained; then, based on the set of physical property parameters of the data point to be judged, the set of physical property parameters of the neighboring data points of the data point to be judged, the coordinate values ​​of the data point to be judged, and the coordinate values ​​of the neighboring data points of the data point to be judged, the gradient vector corresponding to each physical property parameter in the set of physical property parameters of the data point to be judged is obtained. Next, the sum of the magnitudes of the gradient vectors corresponding to all physical property parameters in the set of physical property parameters of the data point to be judged is obtained and recorded as the comprehensive magnitude of the corresponding neighboring data point. Then, the normalized value of the sum of the comprehensive magnitudes of all neighboring data points in the set of neighboring data points of the data point to be judged is obtained and recorded as the second index value of the data point to be judged; finally, the sum of the first index value and the second index value of the data point to be judged is obtained and recorded as the characteristic value of the physical property parameter of the data point to be judged. The specific expression for obtaining the characteristic value of the physical property parameter of the data point to be judged is as follows:

[0029]

[0030] Among them, P A Let A be the eigenvalue of the physical property parameter of the data point to be judged, and let tanh() be the hyperbolic tangent function. M Let A be the number of physical property parameters in the set of physical property parameters for the data point A to be judged. Let A be the difference value of the neighborhood physical property parameters corresponding to the m-th physical property parameter in the set of physical property parameters of the data point A to be judged. n Let Q be the number of neighboring data points in the neighborhood set of the data point A to be judged. n Let be the comprehensive modulus of the nth neighboring data point in the neighborhood set of the data point A to be judged. In this formula, the hyperbolic tangent function is used to normalize the data. Additionally... The larger the value of Q, the more significant the difference between the physical properties of the data point A and its neighboring properties. This indicates a greater likelihood that the data point A belongs to a key region such as a hydrothermal reservoir, fault zone, or hydrothermal channel, or is located on the edge of such a region. n The larger the value, the more drastic the change in the physical properties of the data point A within its neighborhood, thus indicating a greater likelihood that the data point A belongs to a key region such as a thermal reservoir, fault zone, or hydrothermal channel, or is located on the edge of such a region. Because... and Q n When P is larger and larger, A The larger P is, the more likely it is to be. A The larger the value of P, the greater the likelihood that the data point A to be judged belongs to a key area such as a thermal reservoir, fault zone, or hydrothermal channel, or is located on the edge of such a key area. Therefore, the greater the likelihood that data point A will be subsequently identified as a core data point. Conversely, the smaller the value of P... A The smaller the value, the less likely the data point A to belong to a key area such as a thermal reservoir, fault zone, or hydrothermal channel. Therefore, the greater the likelihood that the data point A will be classified as an isolated data point or a noisy data point.

[0031] In this embodiment, the specific process of obtaining the difference value of the neighborhood physical property parameters corresponding to each physical property parameter in the set of physical property parameters of the data point to be judged A is as follows: For the a-th physical property parameter in the set of physical property parameters of the data point to be judged A, in the set of physical property parameters of all neighborhood data points in the set of neighborhood data points of the data point to be judged A, the mean value of all physical property parameters with the same parameter type as the a-th physical property parameter is obtained, and recorded as the neighborhood mean value of the a-th physical property parameter. The absolute value of the difference between the a-th physical property parameter and the neighborhood mean value of the a-th physical property parameter is obtained, and recorded as the difference value of the neighborhood physical property parameter corresponding to the a-th physical property parameter. If the a-th physical property parameter is resistivity, then the mean value of all resistivities in the set of physical property parameters of all neighborhood data points in the set of neighborhood data points of the data point to be judged A is the neighborhood mean value of the a-th physical property parameter.

[0032] In this embodiment, the specific process of obtaining the gradient vector corresponding to each physical property parameter in the physical property parameter set of the data point to be judged A, based on the physical property parameter set of the data point to be judged A, the physical property parameter set of the neighboring data points of the data point to be judged A, the coordinate value of the data point to be judged A, and the coordinate value of the neighboring data points of the data point to be judged A, is as follows: For the b-th physical property parameter in the physical property parameter set of any neighboring data point B in the neighboring set of the data point to be judged A: First, in the physical property parameter set of the data point to be judged A, obtain the physical property parameter with the same parameter type as the b-th physical property parameter. The parameters are denoted as the same type of physical property parameter as the b-th physical property parameter. That is, if the b-th physical property parameter is resistivity, then the resistivity in the set of physical property parameters of the data point A to be judged is the same type of physical property parameter as the b-th physical property parameter. Then, the result of subtracting the b-th physical property parameter from the same type of physical property parameter is obtained and denoted as the first difference value corresponding to the b-th physical property parameter. After that, the coordinate difference set corresponding to the neighboring data point B is obtained. The c-th coordinate difference in the coordinate difference set corresponding to the neighboring data point B is the coordinate value of the c-th dimension of the data point A to be judged minus the number of neighbors. The result of the coordinate value of the c-th dimension of data point B, where c is no greater than 3, means that the number of coordinate differences in the coordinate difference set corresponding to the neighboring data point B is consistent with the dimension of the spatial data point, both being 3. The first coordinate difference in the coordinate difference set corresponding to the neighboring data point B is the result of subtracting the coordinate value of the first dimension of the data point A to be judged from the coordinate value of the first dimension of the neighboring data point B. The second coordinate difference in the coordinate difference set corresponding to the neighboring data point B is the result of subtracting the coordinate value of the second dimension of the data point A to be judged from the coordinate value of the second dimension of the neighboring data point B. The third coordinate difference in the coordinate difference set corresponding to point B is the result of subtracting the coordinate value of the third dimension of the neighboring data point B from the coordinate value of the third dimension of the data point A to be judged. Then, based on the coordinate difference set corresponding to the neighboring data point B and the first difference value, the gradient vector corresponding to the b-th physical property parameter in the set of physical property parameters of the neighboring data point B is constructed. The c-th element of the gradient vector corresponding to the b-th physical property parameter is the ratio of the first difference value to the c-th coordinate difference value in the coordinate difference set corresponding to the neighboring data point B, that is, the gradient vector corresponding to the b-th physical property parameter is... Where D1 is the first difference value, X1 is the first coordinate difference in the set of coordinate differences corresponding to neighboring data point B, Y2 is the second coordinate difference in the set of coordinate differences corresponding to neighboring data point B, and Z3 is the third coordinate difference in the set of coordinate differences corresponding to neighboring data point B; in addition, the magnitude of the gradient vector corresponding to the b-th physical property parameter is... The larger the magnitude of the gradient vector corresponding to the b-th physical property parameter, the more it indicates that the data point A to be judged is in its neighborhood, and the more drastic the change in the physical property parameter of the data point A to be judged.

[0033] Furthermore, the reason for obtaining the physical property parameter characteristic values ​​in this embodiment is that the edge regions of key areas such as geothermal reservoirs, fault zones, and hydrothermal channels are affected by the discontinuity of adjacent rock layers or structures, which can cause sudden changes in the physical property parameters of these key areas. The physical property parameter characteristic values ​​calculated above can reflect whether there is a sudden change in the physical property parameters of the corresponding data point in the neighborhood of the data point to be judged. That is, the larger the physical property parameter characteristic value of the data point to be judged, the greater the possibility of a sudden change in the physical property parameters of the corresponding data point in the neighborhood of the data point to be judged. Therefore, the data point to be judged is more likely to belong to key areas such as geothermal reservoirs, fault zones, and hydrothermal channels. In order to classify the data points in the edge regions of key areas into clusters related to key areas as much as possible, this embodiment needs to further determine the core data points in the data points to be judged based on the magnitude of the physical property parameter characteristic values.

[0034] After obtaining the physical property parameter feature values ​​of each data point to be judged, in order to further classify the data points in the edge areas of the key region into clusters related to the key region, this embodiment will next obtain the structural distribution feature value of the data point to be judged based on the distribution characteristics of the spatial data points within a preset neighborhood radius. The structural distribution feature value is also one of the key parameters for subsequently obtaining the core data points in the data point to be judged. The specific process for obtaining the structural distribution feature value of the data point to be judged is as follows:

[0035] First, the center data point within the preset neighborhood radius of each data point to be judged is obtained and denoted as the neighborhood center data point of the corresponding data point to be judged. That is, the coordinate value of the r-th dimension of the neighborhood center data point of any data point to be judged is the mean of the coordinate values ​​of the r-th dimension of all neighborhood data points in the neighborhood set of the data point to be judged, where r is not greater than 3. Then, based on the distance between the data point to be judged and each neighborhood data point in the neighborhood set of the corresponding data point to be judged, the neighborhood concentration feature value and neighborhood uniformity feature value of the data point to be judged are obtained. The sum of the neighborhood concentration feature value and the neighborhood uniformity feature value of the data point to be judged is denoted as the structural distribution feature value of the corresponding data point to be judged. That is, the structural distribution feature value of any data point to be judged is the sum of the neighborhood concentration feature value and the neighborhood uniformity feature value of the data point to be judged.

[0036] In this embodiment, the specific process of obtaining the neighborhood concentration feature value and neighborhood uniformity feature value of the data point to be judged based on the distance between the data point to be judged and each neighboring data point in the corresponding neighborhood set is as follows: For any data point A to be judged:

[0037] First, obtain the Euclidean distance between the data point to be judged A and the center data points of its neighborhood, and denote it as the feature distance of the data point to be judged A. Then, obtain the Euclidean distance between each neighboring data point in the neighborhood set of the data point to be judged A and its center data point, and denote it as the feature distance of the corresponding neighboring data point. Next, obtain the mean of the feature distances of all neighboring data points in the neighborhood set of the data point to be judged A, and denote it as the feature mean of the data point to be judged A. Finally, obtain the normalized value obtained by adding the feature distances of the data point to be judged A and its feature mean, and denote it as the neighborhood concentration feature value of the data point to be judged A. The neighborhood concentration feature value of the data point to be judged A is... Where, d A Let A be the feature distance of the data point A to be judged. n Let d be the number of neighboring data points in the neighborhood set of the data point A to be judged. n The feature distance is the nth neighboring data point in the neighborhood set of the data point A to be judged. The hyperbolic tangent function here is also used for data normalization; and when d(C i ,F i The larger and A larger value indicates a less compact distribution of data points within the neighborhood of data point A, or a more scattered distribution. In geological exploration, a more scattered distribution of data points within the neighborhood increases the likelihood that the corresponding data point belongs to a key area such as a geothermal reservoir, fault zone, or hydrothermal channel, or is located on the edge of such a key area. Therefore, a larger neighborhood concentration characteristic value for data point A indicates a greater likelihood that data point A belongs to a key area such as a geothermal reservoir, fault zone, or hydrothermal channel, or is located on the edge of such a key area. Consequently, the likelihood of subsequently identifying data point A as a core data point increases.

[0038] Next, obtain the absolute value of the difference between the feature distance of each neighboring data point in the neighborhood set of the data point to be judged and the feature mean of the data point to be judged, and record it as the difference to be analyzed for the corresponding neighboring data point. Then, obtain the standard deviation of the difference to be analyzed for all neighboring data points in the neighborhood set of the data point to be judged, and record it as the neighborhood uniformity feature value of the data point to be judged. Furthermore, the larger the neighborhood uniformity characteristic value of the data point A to be judged, the more uneven the distribution of data points in the neighborhood of the data point A to be judged is, and the more scattered the distribution of data points in the neighborhood of the data point A to be judged is. Key areas such as thermal reservoirs, fault zones, and hydrothermal channels usually exhibit irregular expansion, resulting in a scattered distribution of data points in the edge areas of these key areas. Therefore, the larger the neighborhood uniformity characteristic value of the data point A to be judged, the greater the probability that the data point A to be judged belongs to key areas such as thermal reservoirs, fault zones, and hydrothermal channels, or belongs to the edge areas of key areas such as thermal reservoirs, fault zones, and hydrothermal channels. In this case, the greater the probability that the data point A to be judged will be identified as a core data point in the future.

[0039] Furthermore, the reason for obtaining the structural distribution feature value in this embodiment is that the edge regions of key areas such as geothermal reservoirs, fault zones, and hydrothermal channels have many lithological transition zones and structural features such as tectonic fracture zones and stratigraphic faults. These structural features make the data point distribution in the edge regions of key areas such as geothermal reservoirs, fault zones, and hydrothermal channels relatively scattered. Therefore, this embodiment analyzes the scattered distribution of data points in the neighborhood of the data point to be judged to reflect the probability that the corresponding data point belongs to key areas such as geothermal reservoirs, fault zones, and hydrothermal channels, or to the edge regions of key areas such as geothermal reservoirs, fault zones, and hydrothermal channels. Moreover, the larger the structural distribution feature value of the data point to be judged, the more scattered the distribution of data points in the neighborhood of the corresponding data point to be judged, and the greater the probability that the corresponding data point to be judged belongs to key areas such as geothermal reservoirs, fault zones, and hydrothermal channels. In order to classify the data points in the edge regions of key areas into clusters related to key areas as much as possible, this embodiment needs to further determine the core data points in the data points to be judged based on the magnitude of the structural distribution feature value.

[0040] Therefore, this embodiment obtains the physical property characteristic values ​​and structural distribution characteristic values ​​of the data points to be judged through the above process. Since both the physical property characteristic values ​​and structural distribution characteristic values ​​of the data points to be judged can reflect the possibility that the corresponding data points belong to key areas such as thermal reservoirs, fault zones, and hydrothermal channels, or to the edge areas of key areas such as thermal reservoirs, fault zones, and hydrothermal channels, this embodiment, after obtaining the physical property characteristic values ​​and structural distribution characteristic values ​​of the data points to be judged, combines the physical property characteristic values ​​and structural distribution characteristic values ​​of the data points to be judged to obtain the comprehensive density attribute characteristic value corresponding to the data points to be judged. Subsequently, the comprehensive density attribute characteristic value corresponding to the data points to be judged is used to obtain the data points to be judged. The second data point in the set of data points is also a core data point. In this embodiment, based on the physical property parameter characteristic value and structural distribution characteristic value of the data point to be judged, the comprehensive density attribute characteristic value corresponding to the data point to be judged is obtained. The specific process of obtaining the second data point in the set of data points to be judged based on the comprehensive density attribute characteristic value is as follows: First, the sum of the physical property parameter characteristic value of the data point to be judged and the corresponding distribution characteristic value of the data point to be judged is used as the comprehensive density attribute characteristic value corresponding to the data point to be judged. Then, the comprehensive density attribute characteristic value of each neighboring data point in the neighborhood set of the data point to be judged is obtained, and the comprehensive density attribute characteristic value of all neighboring data points in the neighborhood set of the data point to be judged is calculated. The normalized result obtained by accumulating the density attribute feature values ​​is denoted as the target feature value of the corresponding data point to be judged. The larger the target feature value of the data point to be judged, the greater the probability that the corresponding data point belongs to a key area such as a thermal reservoir, fault zone, or hydrothermal channel, or is located in the edge area of ​​a key area such as a thermal reservoir, fault zone, or hydrothermal channel. Therefore, the probability of subsequently identifying the corresponding data point as a core data point is greater. In this embodiment, the hyperbolic tangent function tanh() is used to normalize the sum of the comprehensive density attribute feature values ​​of all neighboring data points in the neighborhood set of the data point to be judged. In addition, since the method for obtaining the comprehensive density attribute feature value of any spatial data point is different from that of the aforementioned data point to be judged, the method for obtaining the comprehensive density attribute feature value of any spatial data point is different from that of the data point to be judged. The method for obtaining the comprehensive density attribute feature values ​​corresponding to the base points is the same, so it will not be described in detail in this embodiment. Then, based on the target feature values ​​of all data points to be judged, a feature value histogram is constructed. The horizontal axis of the feature value histogram is the target feature value, and the vertical axis is the frequency of the corresponding target feature value in the target feature value set. The target feature value set is composed of the target feature values ​​of all data points to be judged. Then, based on each target feature value on the feature value histogram and the frequency of each target feature value, the Otsu algorithm is used to calculate the inter-class variance value corresponding to each target feature value on the feature value histogram, and the target feature value corresponding to the largest inter-class variance value is selected as the judgment index value.Next, all data points whose target feature values ​​are greater than the judgment index value are obtained and recorded as second data points. That is, if the target feature value of any data point is greater than the judgment index value, then that data point is recorded as the second data point. Furthermore, given the known target feature values ​​and their frequencies on the feature value histogram, the process of calculating the inter-class variance corresponding to each target feature value on the feature value histogram using the Otsu algorithm is a well-known technique, therefore it will not be described in detail in this embodiment.

[0041] Since both the first and second data points obtained in this embodiment are core data points, this embodiment obtains the core data points in the spatial data points through the above process. In addition, the reason why the core data points are determined by the target feature value rather than the comprehensive density attribute feature value in this embodiment is as follows: First, the comprehensive density attribute feature value of a single data point only reflects the local anomaly intensity of the corresponding data point location, which is difficult to reflect the overall regional structure. The target feature value, on the other hand, gathers the key information of multiple points in the neighborhood. Only when the comprehensive density attribute feature values ​​of most points in the region are high will a large target feature value appear, thereby further ensuring that the detected second data point is a continuous transition zone or a patch of edge, rather than an isolated point. Second, the data points in the edge or fault area are sparse, and the comprehensive density attribute feature value of a single data point may not be sufficient to reach the judgment threshold. However, by accumulating the comprehensive density attribute feature values, the features of these points can be merged to form a cumulative effect, which can more reliably capture the overall signal of the edge area.

[0042] Clustering module 03 is used to cluster all spatial data points output from the inversion of the target survey area based on all core data points among all spatial data points.

[0043] After obtaining the core data points of all spatial data points output from the inversion of the target survey area, this embodiment can cluster all spatial data points through cluster expansion, boundary point and noise point processing, etc., to obtain various clusters. Furthermore, the processes of obtaining the core data points and performing cluster expansion, boundary point and noise point processing are the same as those in the traditional DBSCAB clustering algorithm after determining the core data points. In other words, compared with the traditional DBSCAB clustering algorithm, this embodiment only changes the core data points. The acquisition process; in addition, cluster expansion refers to expanding the cluster from the unprocessed core data point through density reachability. This includes adding all data points within the preset neighborhood radius of the core data point to the current cluster, and recursively checking whether unvisited points in the neighborhood are core data points. If they are, the neighboring points are expanded to be added to the cluster, and the above process is repeated until the current cluster can no longer be expanded; boundary point and noise point processing includes: if a non-core data point is located within the preset neighborhood radius of a core data point, the non-core data point is assigned to the corresponding cluster; if a non-core data point is not within the preset neighborhood radius of any core data point, the non-core data point is marked as a noise point or an isolated point.

[0044] Furthermore, by changing the method of determining core data points, this embodiment can improve the sensitivity of data points in the edge regions belonging to key areas such as geothermal reservoirs, fault zones, and hydrothermal channels during clustering. This allows data points in the edge regions belonging to key areas such as geothermal reservoirs, fault zones, and hydrothermal channels to be classified into corresponding clusters as much as possible, thereby improving the accuracy of subsequent boundary identification of geothermal reservoirs, fault zones, and hydrothermal channels, as well as the reliability and accuracy of candidate area delineation and target area selection. In other words, by improving the accuracy of classifying data points in the edge regions belonging to key areas such as geothermal reservoirs, fault zones, and hydrothermal channels, this embodiment can improve the accuracy of subsequent analysis or identification of geothermal resources in the survey area.

[0045] Thus, this embodiment completes the clustering of all spatial data points output from the inversion of the target survey area, and also improves the accuracy of the division of data points in the edge areas belonging to key areas such as thermal reservoirs, fault zones, and hydrothermal channels.

[0046] In summary, this embodiment includes a first acquisition module, used to acquire all spatial data points output from the inversion of the target survey area and the set of physical property parameters corresponding to the spatial data points; a second acquisition module, used to obtain a first data point and a data point to be judged from the spatial data points according to the minimum number within a preset neighborhood and a preset neighborhood radius, to obtain the physical property parameter feature value of the data point to be judged according to the set of physical property parameters of all spatial data points within the preset neighborhood radius of the data point to be judged, to obtain the structural distribution feature value of the data point to be judged according to the distribution characteristics of the spatial data points within the preset neighborhood radius of the data point to be judged, to obtain the comprehensive density attribute feature value corresponding to the data point to be judged according to the physical property parameter feature value and the structural distribution feature value, and to obtain a second data point from the data points to be judged according to the comprehensive density attribute feature value corresponding to the data point to be judged, wherein both the first data point and the second data point belong to core data points; and a clustering module, used to cluster all spatial data points output from the inversion of the target survey area according to all core data points among all spatial data points. Furthermore, this embodiment, based on the comprehensive density attribute feature value corresponding to the data point to be judged, can improve the accuracy of dividing data points in the edge areas of key areas such as geothermal reservoirs, fault zones, and hydrothermal channels. Alternatively, this embodiment, based on the comprehensive density attribute feature value corresponding to the data point to be judged, can divide data points in the edge areas of key areas such as geothermal reservoirs, fault zones, and hydrothermal channels into clusters related to key areas such as geothermal reservoirs, fault zones, and hydrothermal channels as much as possible. This can improve the accuracy of subsequent analysis or identification of geothermal resources in the surveyed area. In other words, this embodiment, based on the comprehensive density attribute feature value corresponding to the data point to be judged, can improve the accuracy of clustering.

[0047] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A cloud computing-based deep geothermal resource exploration data processing platform, characterized in that, The deep geothermal resource exploration data processing platform includes: The first acquisition module is used to acquire all spatial data points output by the inversion of the target survey area and the set of physical property parameters corresponding to the spatial data points. The second acquisition module is used to obtain a first data point and a data point to be judged from the spatial data points based on the minimum number within a preset neighborhood and a preset neighborhood radius; to obtain the physical property parameter feature value of the data point to be judged based on the set of physical property parameters of all spatial data points within the preset neighborhood radius of the data point to be judged; to obtain the structural distribution feature value of the data point to be judged based on the distribution characteristics of the spatial data points within the preset neighborhood radius of the data point to be judged; to obtain the comprehensive density attribute feature value corresponding to the data point to be judged based on the physical property parameter feature value and the structural distribution feature value; and to obtain a second data point from the data points to be judged based on the comprehensive density attribute feature value corresponding to the data point to be judged, wherein both the first data point and the second data point belong to core data points. The clustering module is used to cluster all spatial data points output from the inversion of the target survey area based on all core data points among all spatial data points; The method for obtaining the structural distribution feature values ​​of the data points to be judged includes: Obtain the center data point within the preset neighborhood radius of the data point to be judged, and record it as the neighborhood center data point of the corresponding data point to be judged; For any data point to be judged, the distance between the data point to be judged and its neighborhood center data point is denoted as the feature distance of the data point to be judged; the distance between each neighboring data point in the neighborhood set of the data point to be judged and its neighborhood center data point is denoted as the feature distance of the corresponding neighboring data point; the mean of the feature distances of all neighboring data points in the neighborhood set of the data point to be judged is denoted as the feature mean; the normalized value obtained by adding the feature distance of the data point to the feature mean is denoted as the neighborhood concentration feature value of the data point to be judged; the absolute value of the difference between the feature distance of each neighboring data point in the neighborhood set of the data point to be judged and the feature mean is denoted as the difference to be analyzed for the corresponding neighboring data point; the standard deviation of the difference to be analyzed for all neighboring data points in the neighborhood set of the data point to be judged is denoted as the neighborhood evenness feature value of the data point to be judged; and the sum of the neighborhood concentration feature value and the neighborhood evenness feature value of the data point to be judged is denoted as the structural distribution feature value of the data point to be judged.

2. The cloud computing-based deep geothermal resource exploration data processing platform as described in claim 1, characterized in that, The method for obtaining the first data point and the data point to be judged in the spatial data points includes: For any spatial data point, if the number of spatial data points within a preset neighborhood radius of the spatial data point is less than the minimum number within the preset neighborhood, then the spatial data point is recorded as a data point to be judged; otherwise, the spatial data point is recorded as the first data point.

3. The cloud computing-based deep geothermal resource exploration data processing platform as described in claim 1, characterized in that, The method for obtaining the physical property parameter characteristic values ​​of the data points to be judged includes: For any data point to be judged: The set of all spatial data points within a preset neighborhood radius of the data point to be judged is denoted as the neighborhood set of the data point to be judged. All spatial data points in the neighborhood set are denoted as neighborhood data points. Based on the set of physical property parameters of the neighborhood data points, the difference value of the neighborhood physical property parameters corresponding to each physical property parameter in the set of physical property parameters of the data point to be judged is obtained. The normalized value of the sum of the difference values ​​of the neighborhood physical property parameters corresponding to all physical property parameters in the set of physical property parameters of the data point to be judged is used as the first index value of the data point to be judged. Based on the physical property parameters of the data point to be judged... The set of physical property parameters of the neighboring data points, the coordinate values ​​of the data point to be judged, and the coordinate values ​​of the neighboring data points are used to obtain the gradient vector corresponding to each physical property parameter in the set of physical property parameters of the neighboring data points. The sum of the magnitudes of the gradient vectors corresponding to all physical property parameters in the set of physical property parameters of the neighboring data points is recorded as the comprehensive magnitude of the corresponding neighboring data point. The normalized value of the sum of the comprehensive magnitudes of all neighboring data points in the set of neighboring data points is recorded as the second index value of the data point to be judged. The sum of the first index value and the second index value of the data point to be judged is recorded as the physical property parameter feature value of the data point to be judged.

4. The cloud computing-based deep geothermal resource exploration data processing platform as described in claim 3, characterized in that, The method for obtaining the difference values ​​of neighborhood physical parameters corresponding to each physical parameter in the set of physical parameter values ​​of the data points to be judged includes: For the a-th physical property parameter in the set of physical property parameters of the data point to be judged, in the set of physical property parameters of all neighboring data points in the neighborhood set, the mean of all physical property parameters with the same parameter type as the a-th physical property parameter is recorded as the neighborhood mean of the a-th physical property parameter, and the absolute value of the difference between the a-th physical property parameter and the neighborhood mean of the a-th physical property parameter is recorded as the neighborhood difference value of the a-th physical property parameter.

5. The cloud computing-based deep geothermal resource exploration data processing platform as described in claim 3, characterized in that, The method for obtaining the gradient vector corresponding to each physical property parameter in the set of physical property parameters of the neighborhood data points includes: For the b-th physical property parameter in the physical property parameter set of any neighborhood data point in the neighborhood set: Physical property parameters in the physical property parameter set of the data point to be judged that have the same parameter type as the b-th physical property parameter are denoted as physical property parameters of the same type as the b-th physical property parameter. The result of subtracting the b-th physical property parameter from the physical property parameter of the same type is denoted as the first difference value corresponding to the b-th physical property parameter. A coordinate difference set corresponding to the neighborhood data point is obtained. Based on the coordinate difference set and the first difference value, a gradient vector corresponding to the b-th physical property parameter in the physical property parameter set of the neighborhood data point is constructed. The c-th element in the gradient vector corresponding to the b-th physical property parameter is the ratio of the first difference value to the c-th coordinate difference value in the coordinate difference set. The c-th coordinate difference value in the coordinate difference set is the result of subtracting the c-th coordinate value of the neighborhood data point from the c-th dimension coordinate value of the data point to be judged. The spatial data point is a three-dimensional data point.

6. The cloud computing-based deep geothermal resource exploration data processing platform as described in claim 1, characterized in that, The comprehensive density attribute feature value corresponding to the data point to be judged is the sum of the physical property parameter feature value and the structural distribution feature value of the corresponding data point to be judged.

7. The cloud computing-based deep geothermal resource exploration data processing platform as described in claim 3, characterized in that, The method for obtaining the second data point among the data points to be judged includes: The target feature value of the data point to be judged is obtained. The target feature value of the data point to be judged is the normalized value of the result obtained by accumulating the comprehensive density attribute feature values ​​of all neighboring data points in the neighborhood set of the data point to be judged. The method for obtaining the comprehensive density attribute feature value of any data point is the same as the method for obtaining the comprehensive density attribute feature value of the data point to be judged. Based on the target feature values ​​of all data points to be judged, a feature value histogram is constructed. The horizontal axis of the feature value histogram represents the target feature value, and the vertical axis represents the frequency of occurrence of the corresponding target feature value. Based on each target feature value and its frequency of occurrence in the feature value histogram, the inter-class variance value corresponding to each target feature value in the feature value histogram is obtained. The target feature value corresponding to the largest inter-class variance value is selected as the judgment index value. All data points to be judged whose target feature value is greater than the judgment index value are recorded as the second data points.

Images