CN120470953A - A gold mine target area optimization method based on spatially constrained modified embedded geochemical anomaly and its system and storage - Google Patents
A gold mine target area optimization method based on spatially constrained modified embedded geochemical anomaly and its system and storageInfo
- Publication number
- CN120470953A CN120470953A CN202510988508.5A CN202510988508A CN120470953A CN 120470953 A CN120470953 A CN 120470953A CN 202510988508 A CN202510988508 A CN 202510988508A CN 120470953 A CN120470953 A CN 120470953A
- Authority
- CN
- China
- Prior art keywords
- target area
- anomaly
- geochemical
- gold mine
- matrix
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a gold mine target area optimization method based on space constraint modification embedded geochemistry anomaly, a system and a memory thereof. The method comprises the steps of collecting geochemical data of a target area, preprocessing, carrying out spatial serialization standardized data matrix and adjacent matrix, adopting global transducer coding modeling to obtain hidden state sequences of global element interaction characteristics of all sampling points of the target area, sequentially constructing unitary potential energy and binary potential energy, combining CRF to establish a geochemical anomaly area identification model of all sampling points of the target area, carrying out anomaly scoring and marking on the geochemical anomaly area marked by the model, and carrying out comprehensive evaluation to finish gold mine target area optimization. The system containing the computer program memory provided by the method can effectively solve the problems of large calculated amount and low accuracy in the prior art, and the area under the ROC curve of the optimization process of the gold mine target area by adopting the system can reach 0.92 through test, so that the system has remarkable advantages.
Description
Technical Field
The invention relates to a gold mine target area optimization method, in particular to a gold mine target area optimization method based on space constraint modification embedded geochemistry anomaly, a system and a memory thereof.
Background
In gold exploration work, the common general knowledge of geochemistry is the key element in determining potential gold target areas. For a long time, the technology of identifying the geochemical anomaly is continuously developed, but the traditional method still has obvious defects.
Early, the mean-standard deviation method was the usual means of geochemical anomaly identification. The method is simple to operate, and the abnormality is identified by only calculating the mean value and the standard deviation of the geochemical index and taking the standard deviation of a certain multiple as a threshold value. However, this approach is too simplistic and rough. In a practical gold exploration scenario, geochemical data is complex and multidimensional, often containing multi-dimensional aqueous sediment data such as the content of various elements such as gold, silver, copper, lead, zinc, and other relevant geological parameters. The mean-standard deviation method can only analyze single or few geochemical indexes, can not comprehensively utilize multidimensional data information, is difficult to perceive for geochemical anomaly modes formed by complex combination of various elements, and causes a large number of potential gold mine target areas to be missed.
With the development of machine learning technology, algorithms such as support vector machines, decision trees, and the like are also applied to geochemical anomaly recognition. These algorithms perform well when processing simple data, but the dimension disaster problem stands out when faced with high-dimensional geochemical data. The high dimensional data implies more features, which not only increases computational complexity, but also tends to result in model overfitting, making the model less generalizable on new data. The geochemical data has time and space sequence characteristics, such as content change of geochemical elements in the same region at different times, distribution relation of geochemical elements in different space positions, and the like. And the traditional machine learning algorithms such as a support vector machine, a decision tree and the like are difficult to effectively mine and utilize the long-term dependency relationship and the sequence information, so that geochemical anomalies related to gold ores cannot be accurately identified.
In addition, the spatial distribution of geochemical data is not completely random, but rather there is some spatial autocorrelation. The existing method often does not fully consider the spatial autocorrelation, so that the actual geological condition cannot be accurately reflected when the geochemical anomaly and the preferential gold target area are identified, the finally determined gold target area is poor in accuracy, and the efficiency and cost of the subsequent exploration work are affected.
In summary, in the prior art, when processing high-dimensional geochemical data, there is a serious disadvantage in fully mining data characteristics, utilizing sequence information, accurately identifying geochemical anomalies, and the like, and a new method is needed to improve the accuracy and efficiency of gold target region optimization.
Disclosure of Invention
Aiming at the problems existing in the prior art, the first object of the invention is to provide a gold mine target area optimization method based on space constraint modification embedded geochemical anomalies. The method fully and deeply excavates the characteristics and the relations in the geochemical data, aims to solve various problems existing in the prior art when the high-dimensional geochemical data is processed, remarkably improves the accuracy and the efficiency of gold target area optimization, and provides more scientific and reliable technical support for gold exploration.
A second object of the present invention is to provide a computer readable memory having embodied thereon a computer program for implementing the above uniformity control method, which can be read and executed.
The third object of the invention is to provide a gold mine target area optimization system based on space constraint modification embedded geochemistry anomalies. The system can realize accurate identification and optimization of the gold mine target area based on the excellent accuracy and efficiency of the method, the process has extremely small calculation force, the equipment cost is greatly reduced only by a conventional computer, and the area under the ROC curve of the optimization process of the gold mine target area by adopting the system can reach 0.92 through test, so that the system has remarkable advantages.
In order to achieve the technical aim, the invention provides a gold mine target area optimization method based on space constraint modification embedded geochemistry anomaly, which comprises the following steps:
S1, collecting geochemical data of a target area, and carrying out spatial serialization standardized data matrix and adjacent matrix after pretreatment;
S2, carrying out global transform coding modeling by adopting the preprocessed data to obtain a hidden state sequence containing global element interaction characteristics of each sampling point of a target area;
step S3, according to the hidden state sequence obtained in the step S2, respectively constructing unitary potential energy and binary potential energy by mapping hidden states and fusing geographic distances and element similarity through a full connection layer, and then establishing a geochemical anomaly region identification model of each sampling point of a target region through a CRF;
And S4, carrying out abnormal scoring on the geochemical abnormal region marked by the model obtained in the step S3 through the autoregressive characteristic of the transducer model, marking the geochemical abnormal region higher than a scoring threshold as an abnormal region, and carrying out comprehensive evaluation to finish gold mine target region optimization.
As a preferred embodiment, the geochemical data includes spatial coordinates of each sampling point, an element type of each sampling point, and a concentration value of the element type.
As a preferable scheme, the pretreatment process is multistage pretreatment, comprising a cleaning process, a standardization process and an adjacency matrix construction process, and the specific process is as follows:
step S1-1, cleaning, namely after cleaning the data, filling the estimated value of the missing element by adopting a K Nearest Neighbor (KNN) interpolation method combined with the spatial position, wherein the estimated value of the i-th missing element is as follows:
formula 1: ;
Formula 2: ;
step S1-2, standardization processing, namely, performing standardization processing on the feature matrix Normalizing or normalizing to eliminate dimension difference and obtain normalized characteristic matrix;
Step S1-3, adjacency matrix construction processing, namely constructing Voronoi adjacency matrix based on Delaunay triangulationIf and only if the two points are Voronoi adjacencies,Providing a topological structure for space constraint of a subsequent CRF;
In the formula (1) and (2), Is thatK spatial nearest neighbors of a point; Concentration values for elemental species; Is the weight; Is that Longitude of the point space location; Is that Dimension of the point space position; the influence degree of the space distance on the weight is controlled as the bandwidth parameter.
The global transform coding modeling process is that a preprocessed and serialized standardized data matrix is input into a transform coder with global perceptibility constructed by a plurality of layers of improved Transformer Block, so as to obtain a hidden state sequence containing the global element interaction characteristics of each sampling point of a target area.
The mechanism of the encoder allows each position to interact with all positions in the sequence, breaks the limitation of the traditional autoregressive model, can comprehensively capture complex relations among elements in data, and can mine hidden nonlinear relations in the earth data in such a way so as to provide rich and valuable characteristic information for subsequent analysis.
As a preferred embodiment, the multilayer modified Transformer Block thThe calculation flow of the layer is as follows:
The attention mechanism of the global coverage is adopted, the limitation of the mask matrix M is removed, each position is allowed to freely interact with all positions in the sequence, and the calculation formula is as follows:
formula 3: ;
Formula 4: ;
formula 5: ;
Formula 6: ;
In the 3-6 of the present invention, Is a standardized data matrix; is a polling matrix; is a key matrix; is a value matrix; is a learnable parameter.
The attention mechanism adopted by the invention can lead the model to comprehensively capture the complex relationship among elements in the geochemical data, break through the limitation that the traditional autoregressive model can only locally interact, and mine the hidden nonlinear relationship among the elements. For example, in a complex geochemical environment, the relationship between certain trace elements and gold ores may be ignored by the traditional model, but through the attention mechanism of global interaction, potential links between the trace elements and the gold ores can be found, so that richer characteristic information is provided for accurately identifying the geochemical anomalies subsequently.
As a preferable scheme, the construction process of the geochemical anomaly area identification model is as follows:
Step S3-1, outputting hidden state sequence by using a transform encoder Mapping the hidden state into initial abnormal probability, namely unitary potential energy, through a full connection layer;
s3-2, fusing the geographic distance and the element similarity to construct a dynamic transfer matrix, namely binary potential energy;
and step S3-3, optimizing the unitary potential energy and the binary potential energy through global constraint of the CRF, and solving the maximum posterior probability to obtain the binary potential energy.
As a preferable scheme, the calculation process of the unitary potential energy is as follows:
formula 7: ;
The binary potential energy is calculated by the following steps:
Formula 8: ;
the formula for solving the maximum posterior probability is as follows:
Formula 9: ;
In the 8-9 type, the following, The global element interaction feature of the ith sampling point is contained; to learn parameters, corresponding labels ;In order to bias the term vector,;Is a Voronoi adjacency matrix; Is the geographic distance; is the element similarity attenuation coefficient; For all sets of contiguous edges, Is a normalization factor.
The binary potential energy calculation process fully considers the spatial characteristics of geochemical data, and the probability of state transition between adjacent sampling points is measured by using the geographic distance and the element similarity. For example, spatially adjacent sampling points with similar element concentrations, which have a higher probability of having the same state (normal or abnormal), can in this way better utilize the spatial correlation of the data for anomaly identification. Further, in the process of solving the maximum posterior probability, on the basis of considering the state probability (unitary potential energy) of each sampling point and the state transition probability (binary potential energy) between adjacent sampling points, the most probable state of each sampling point, namely whether the most probable state is a geochemical anomaly region or not, is determined through CRF global optimization, so that the geochemical anomaly is accurately identified.
The process of carrying out anomaly scoring on the geochemical anomaly area is that the input high-dimensional localization data is reconstructed after being processed by the step S2 and the step S3 by utilizing the autoregressive characteristic of a transducer model, and the difference between the original input data and the reconstructed data is calculated, namely the anomaly score is calculated by the following calculation process:
Formula 10: ;
In the process of 10, the process is carried out, Represent the firstSample point numberOriginal input values of the individual elements; Represent the first Sample point numberPredicted values of the individual elements.
As a preferred embodiment, the scoring threshold is determined according to the da You log index, and an anomaly region is marked when the anomaly score is above the threshold.
As a preferred scheme, the key factors of the comprehensive evaluation are the geological structure stability of the target area and the spatial relationship of known gold ores.
The invention also provides a storable computer, which comprises a computer program, wherein the computer program can realize the gold mine target area optimization method based on space constraint modification embedded geochemical anomalies.
The invention also provides a gold mine target area optimization system based on space constraint modification embedded geochemistry anomaly, which comprises a programmable logic controller, a central processing unit and the storable computer, wherein the programmable logic controller collects production process data, inputs the production process data to the central processing unit and executes a computer program on the storable computer to obtain the gold mine target area optimization system.
Compared with the prior art, the technical scheme of the invention has the beneficial technical effects that:
1) The method provided by the invention fully exerts the capturing capability of a self-attention mechanism on long-range dependency relationship in long-sequence data by introducing a transducer model, excavates nonlinear relationship among elements in high-dimensional geochemical data, combines a conditional random field CRF, utilizes modeling capability of the nonlinear relationship on data context, particularly considers the space characteristics of the geochemical data, so as to more accurately identify geochemical anomalies, and realizes gold mine target region optimization by reconstructing error anomaly scoring and comprehensive evaluation, thereby greatly improving accuracy and efficiency of optimization results.
2) According to the technical scheme provided by the invention, based on the excellent accuracy and efficiency of the method, the accurate identification and optimization of the gold mine target area can be realized, the required calculation force in the process is extremely small, the equipment cost can be greatly reduced only by a conventional computer, and the area under the ROC curve in the optimization process of the gold mine target area by adopting the system can reach 0.92 through test, so that the remarkable advantage is shown.
Drawings
FIG. 1 is a graph of ROC curves for the methods provided in example 1 and comparative example 1 of the present invention;
fig. 2 is a graph of ROC curves for the methods provided by example 1 and comparative examples 2 and 3 of the present invention.
Detailed Description
The present invention will be described more fully hereinafter with reference to the accompanying drawings, in order to facilitate an understanding of the invention. It should be noted that the described embodiments are only some embodiments, not all embodiments, of the present invention. All other embodiments, which can be obtained by a person skilled in the art without creative efforts, are within the protection scope of the present invention based on the embodiments of the present invention.
Example 1
The embodiment provides a gold mine target area optimization method based on space constraint modification embedded geochemistry anomaly, which selects a Jian-xi-North gold mine collection area in Shandong province as a research area and collects geochemistry measured values of 39 element water system sediments in the area as original data. The deposit verification set is known to contain 27 large and medium size deposits. In the data processing and model running process, the high efficiency and accuracy of calculation are ensured by means of the strong calculation capability of the Nvidia RTX 3090Ti 128G display card, and the specific process is as follows:
S1, collecting geochemical data of a target area, and carrying out spatial serialization standardized data matrix and adjacent matrix after pretreatment;
Based on the regional water detection sediment data, the input data comprises the geographic coordinates (longitude) of the sampling points Latitude and longitudeI=1, once again, N) and 39 concentration values of elements (Au, as, cu, etc.)(D=1,....39) performing multistage preprocessing on the data, including a cleaning process, a normalization process and an adjacency matrix construction process, wherein the specific processes are as follows:
step S1-1, cleaning, namely after cleaning the data, filling the estimated value of the missing element by adopting a K Nearest Neighbor (KNN) interpolation method combined with the spatial position, wherein the estimated value of the i-th missing element is as follows:
formula 1: ;
Formula 2: ;
step S1-2, standardization processing, namely, performing standardization processing on the feature matrix Normalizing or normalizing to eliminate dimension difference and obtain normalized characteristic matrix;
Step S1-3, adjacency matrix construction processing, namely constructing Voronoi adjacency matrix based on Delaunay triangulationIf and only if the two points are Voronoi adjacencies,Providing a topological structure for space constraint of a subsequent CRF;
In the formula (1) and (2), Is thatK spatial nearest neighbors of a point; Concentration values for elemental species; Is the weight; Is that Longitude of the point space location; Is that Dimension of the point space position; the influence degree of the space distance on the weight is controlled as the bandwidth parameter.
S2, carrying out global transform coding modeling by adopting the preprocessed data to obtain a hidden state sequence containing global element interaction characteristics of each sampling point of a target area;
The global transform coding modeling process comprises the steps of inputting a preprocessed and serialized standardized data matrix into a transform coder with global perceptibility constructed by a plurality of layers of improved Transformer Block to obtain a hidden state sequence containing global element interaction characteristics of each sampling point of a target area, wherein the third layer of improved Transformer Block The calculation flow of the layer is as follows:
The attention mechanism of the global coverage is adopted, the limitation of the mask matrix M is removed, each position is allowed to freely interact with all positions in the sequence, and the calculation formula is as follows:
formula 3: ;
Formula 4: ;
formula 5: ;
Formula 6: ;
In the 3-6 of the present invention, Is a standardized data matrix; is a polling matrix; is a key matrix; is a value matrix; is a learnable parameter.
Step S3, according to the hidden state sequence obtained in the step S2, respectively constructing unitary potential energy and binary potential energy by mapping hidden states and fusing geographic distances and element similarity through a full connection layer, and then establishing a geochemical anomaly region identification model of each sampling point of a target region through a CRF;
the construction process of the geochemical anomaly area identification model comprises the following steps:
Step S3-1, outputting hidden state sequence by using a transform encoder Mapping the hidden state into initial abnormal probability, namely unitary potential energy, through a full connection layer;
s3-2, fusing the geographic distance and the element similarity to construct a dynamic transfer matrix, namely binary potential energy;
and step S3-3, optimizing the unitary potential energy and the binary potential energy through global constraint of the CRF, and solving the maximum posterior probability to obtain the binary potential energy.
As a preferable scheme, the calculation process of the unitary potential energy is as follows:
formula 7: ;
The binary potential energy is calculated by the following steps:
Formula 8: ;
the formula for solving the maximum posterior probability is as follows:
Formula 9: ;
In the 8-9 type, the following, The global element interaction feature of the ith sampling point is contained; to learn parameters, corresponding labels ;In order to bias the term vector,;Is a Voronoi adjacency matrix; Is the geographic distance; is the element similarity attenuation coefficient; For all sets of contiguous edges, Is a normalization factor.
S4, performing abnormal scoring on the geochemical abnormal region marked by the model obtained in the step S3 through the autoregressive characteristic of the transducer model, marking the geochemical abnormal region higher than a scoring threshold as an abnormal region, and performing comprehensive evaluation to finish gold mine target region optimization;
the process of carrying out anomaly scoring on the geochemical anomaly region comprises the steps of reconstructing input high-dimensional localization data after processing in the step S2 and the step S3 by utilizing the autoregressive characteristic of a transducer model, and calculating the difference between the original input data and the reconstructed data, namely the anomaly score, wherein the calculation process comprises the following steps:
Formula 10: ;
In the process of 10, the process is carried out, Represent the firstSample point numberOriginal input values of the individual elements; Represent the first Sample point numberPredicted values of the individual elements.
As a preferred embodiment, the scoring threshold is determined according to the da You log index, and an anomaly region is marked when the anomaly score is above the threshold.
As a preferred scheme, the key factors of the comprehensive evaluation are the geological structure stability of the target area and the spatial relationship of known gold ores.
The greater the reconstruction error, the greater the degree of deviation of the geochemical data pattern from the normal pattern, i.e., the greater the degree of geochemical anomaly, for that region. Based on the reconstruction error, an anomaly score is further calculated, and a region with a score higher than a certain threshold is determined as a geochemical anomaly region. The abnormal score can be more objectively and accurately estimated by the mode of calculating the abnormal score through the reconstruction error, and subjectivity and limitation of manually setting the threshold are avoided.
And calculating the obtained abnormal score according to the trained CRF model. An anomaly score threshold is set based on the da You log index. And then marking the area corresponding to the data point with the abnormality score being larger than the threshold value and the CRF judging as the abnormality label as a potential gold mine target area. Further comprehensive evaluation is carried out on the potential target areas, such as the factors of geographic positions, surrounding geological structures and the like of the potential target areas are considered, and finally, the gold mine target areas with preferential exploration are determined.
The number of layers of the transducer encoder adopted by the invention is 8, the attention head number is 2, the dimension of the hidden layer is 512, the dropout proportion is 0.1, the optimizer AdamW, the batch size is 64, and the training round is 100 rounds. The regularization coefficient of CRF is 10e -4, preventing the transfer matrix from overfitting.
The embodiment also provides a storable computer, which comprises a computer program, wherein the computer program can realize the gold mine target area optimization method based on space constraint modification embedded geochemical anomalies.
The embodiment also provides a gold mine target area optimization system based on space constraint modification embedded geochemistry anomaly, which comprises a programmable logic controller, a central processing unit and the storable computer, wherein the programmable logic controller collects production process data, inputs the production process data to the central processing unit and executes a computer program on the storable computer to obtain the gold mine target area optimization system.
In order to better illustrate the excellent effect of the technical scheme provided by the invention, the invention introduces AUC and Morna's I evaluation indexes to perform gold mine target region optimization score on the embodiment 1, wherein Moran's I index is a spatial index for evaluating the spatial aggregation of predicted abnormalities. Moran's I index takes on a value ranging from-1 to 1, and is larger than 0 to indicate positive correlation, namely aggregation phenomenon exists in space, 0 is equal to random distribution, and 0 is smaller than 0 to indicate negative correlation, namely discrete distribution in space. The distribution characteristics of geochemical anomalies in space are further known by calculating Moran's I index, and the capturing capacity of the method for the geochemical anomaly space mode is verified, and the preferable scoring result of the gold mine target area is shown in table 1.
Comparative example 1
This comparative example is identical to example 1, except that Principal Component Analysis (PCA) + local singularity index (Local Singularity Index) is selected for geochemical anomaly identification.
Comparative example 2
This comparative example is identical to example 1 except that an ablation experiment is performed, leaving only the transducer model.
Comparative example 3
This comparative example is identical to example 1, except that an ablation experiment is performed, leaving only the CRF model.
The results of the comparison of the AUC values and Morna's I values of examples 1 and comparative examples 1 to 3 are shown in Table 2.
The comparison experiment and the result show that the method provided by the invention is excellent in AUC value, which shows that the method has remarkable advantages in the aspect of identifying the correlation between geochemical anomalies and mine points, and meanwhile, the Moran's I index result shows that the method has excellent performance in the aspect of capturing the space aggregation of the geochemical anomalies and the preferable gold mine target area. And the feasibility and effectiveness of the invention in the process of optimizing the gold mine target area are fully proved by combining various indexes and comparison experiment results, and the invention has superiority compared with the traditional method and single model.
Claims (10)
1. A gold mine target area optimization method based on space constraint modification embedded geochemistry anomaly, which is characterized by comprising the following steps:
S1, collecting geochemical data of a target area, and carrying out spatial serialization standardized data matrix and adjacent matrix after pretreatment;
S2, carrying out global transform coding modeling by adopting the preprocessed data to obtain a hidden state sequence containing global element interaction characteristics of each sampling point of a target area;
step S3, according to the hidden state sequence obtained in the step S2, respectively constructing unitary potential energy and binary potential energy by mapping hidden states and fusing geographic distances and element similarity through a full connection layer, and then establishing a geochemical anomaly region identification model of each sampling point of a target region through a CRF;
And S4, carrying out abnormal scoring on the geochemical abnormal region marked by the model obtained in the step S3 through the autoregressive characteristic of the transducer model, marking the geochemical abnormal region higher than a scoring threshold as an abnormal region, and carrying out comprehensive evaluation to finish gold mine target region optimization.
2. The method for modifying an embedded geochemical anomaly based gold target area according to claim 1, wherein the geochemical data comprises spatial coordinates of each sampling point, element types of each sampling point and concentration values of the element types.
3. A method for modifying a gold mine target region based on space constraint modification embedded geochemical anomalies according to claim 2, characterized in that:
the pretreatment process is multistage pretreatment, comprising cleaning treatment, standardization treatment and adjacency matrix construction treatment, and comprises the following specific processes:
step S1-1, cleaning, namely after cleaning the data, filling the estimated value of the missing element by adopting a K Nearest Neighbor (KNN) interpolation method combined with the spatial position, wherein the estimated value of the i-th missing element is as follows:
formula 1: ;
Formula 2: ;
step S1-2, standardization processing, namely, performing standardization processing on the feature matrix Normalizing or normalizing to eliminate dimension difference and obtain normalized characteristic matrix;
Step S1-3, adjacency matrix construction processing, namely constructing Voronoi adjacency matrix based on Delaunay triangulationIf and only if the two points are Voronoi adjacencies,Providing a topological structure for space constraint of a subsequent CRF;
In the formula (1) and (2), Is thatK spatial nearest neighbors of a point; Concentration values for elemental species; Is the weight; Is that Longitude of the point space location; Is that Dimension of the point space position; the influence degree of the space distance on the weight is controlled as the bandwidth parameter.
4. The method for optimizing gold mine target area based on space constraint modification embedded geochemistry anomaly, which is characterized in that the global transform coding modeling process comprises the steps of inputting a preprocessed and serialized standardized data matrix into a transform coder with global perceptibility constructed by a plurality of layers of improved Transformer Block to obtain a hidden state sequence containing global element interaction characteristics of each sampling point of a target area;
the first of the multilayer modified Transformer Block The calculation flow of the layer is as follows:
The attention mechanism of the global coverage is adopted, the limitation of the mask matrix M is removed, each position is allowed to freely interact with all positions in the sequence, and the calculation formula is as follows:
formula 3: ;
Formula 4: ;
formula 5: ;
Formula 6: ;
In the 3-6 of the present invention, Is a standardized data matrix; is a polling matrix; is a key matrix; is a value matrix; is a learnable parameter.
5. The method for optimizing gold mine target area based on space constraint modification embedded geochemical anomaly, which is characterized in that the geochemical anomaly area identification model is constructed by the following steps:
Step S3-1, outputting hidden state sequence by using a transform encoder Mapping the hidden state into initial abnormal probability, namely unitary potential energy, through a full connection layer;
s3-2, fusing the geographic distance and the element similarity to construct a dynamic transfer matrix, namely binary potential energy;
and step S3-3, optimizing the unitary potential energy and the binary potential energy through global constraint of the CRF, and solving the maximum posterior probability to obtain the binary potential energy.
6. The method for optimizing gold mine target area based on space constraint modification embedded geochemical anomalies, as claimed in claim 5, is characterized in that:
The calculation process of the unitary potential energy comprises the following steps:
formula 7: ;
The binary potential energy is calculated by the following steps:
Formula 8: ;
the formula for solving the maximum posterior probability is as follows:
Formula 9: ;
In the 8-9 type, the following, The global element interaction feature of the ith sampling point is contained; to learn parameters, corresponding labels ;Is a bias vector;Is a Voronoi adjacency matrix; Is the geographic distance; is the element similarity attenuation coefficient; For all sets of contiguous edges, Is a normalization factor.
7. The gold mine target region optimization method based on space constraint modification embedded geochemistry anomaly, which is characterized in that anomaly scoring of the geochemistry anomaly region is carried out by reconstructing input high-dimensional localization data after processing in step S2 and step S3 by utilizing autoregressive characteristics of a transducer model, and calculating differences between original input data and reconstructed data, namely anomaly score, wherein the calculation process is as follows:
Formula 10: ;
In the process of 10, the process is carried out, Represent the firstSample point numberOriginal input values of the individual elements; Represent the first Sample point numberPredicted values of the individual elements.
8. The method for modifying a gold mine target area with embedded geochemical anomalies based on spatial constraints as set forth in claim 7, wherein the scoring threshold is determined according to a da You log index, and the anomaly score is marked as an anomaly area when the anomaly score is higher than the threshold, and the key factors of the comprehensive evaluation are the geological structure stability of a target area and the spatial relationship of known gold mine.
9. A storable calculator comprises a computer program and is characterized in that the computer program can realize the gold mine target area optimization method based on space constraint modification embedded geochemistry anomaly according to any one of claims 1-8.
10. A gold mine target area optimization system based on space constraint modification embedded geochemistry anomaly is characterized by comprising a programmable logic controller, a central processing unit and a storable computer according to claim 9, wherein the programmable logic controller collects production process data, inputs the production process data to the central processing unit and executes a computer program on the storable computer to obtain the gold mine target area optimization system.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510988508.5A CN120470953A (en) | 2025-07-17 | 2025-07-17 | A gold mine target area optimization method based on spatially constrained modified embedded geochemical anomaly and its system and storage |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510988508.5A CN120470953A (en) | 2025-07-17 | 2025-07-17 | A gold mine target area optimization method based on spatially constrained modified embedded geochemical anomaly and its system and storage |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN120470953A true CN120470953A (en) | 2025-08-12 |
Family
ID=96635387
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202510988508.5A Pending CN120470953A (en) | 2025-07-17 | 2025-07-17 | A gold mine target area optimization method based on spatially constrained modified embedded geochemical anomaly and its system and storage |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN120470953A (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6430547B1 (en) * | 1999-09-22 | 2002-08-06 | International Business Machines Corporation | Method and system for integrating spatial analysis and data mining analysis to ascertain relationships between collected samples and geology with remotely sensed data |
| RU2767159C1 (en) * | 2020-10-02 | 2022-03-16 | Акционерное общество "Сибирский научно-исследовательский институт геологии, геофизики и минерального сырья" | Method for searching for gold-ore and gold-bearing deposits by ore and geochemical associations |
| US20240310554A1 (en) * | 2023-03-16 | 2024-09-19 | Tibet Julong Copper Co., Ltd. | Method for metallogenic prediction by using multi-source heterogeneous information |
| CN119903897A (en) * | 2025-04-02 | 2025-04-29 | 中南大学 | A method and device for selecting gold mine target areas based on element geochemical anomalies |
-
2025
- 2025-07-17 CN CN202510988508.5A patent/CN120470953A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6430547B1 (en) * | 1999-09-22 | 2002-08-06 | International Business Machines Corporation | Method and system for integrating spatial analysis and data mining analysis to ascertain relationships between collected samples and geology with remotely sensed data |
| RU2767159C1 (en) * | 2020-10-02 | 2022-03-16 | Акционерное общество "Сибирский научно-исследовательский институт геологии, геофизики и минерального сырья" | Method for searching for gold-ore and gold-bearing deposits by ore and geochemical associations |
| US20240310554A1 (en) * | 2023-03-16 | 2024-09-19 | Tibet Julong Copper Co., Ltd. | Method for metallogenic prediction by using multi-source heterogeneous information |
| CN119903897A (en) * | 2025-04-02 | 2025-04-29 | 中南大学 | A method and device for selecting gold mine target areas based on element geochemical anomalies |
Non-Patent Citations (2)
| Title |
|---|
| SHUYAN YU ET AL.: "Identification of Geochemical Anomalies Using an End-to-End Transformer", 《NATURAL RESOURCES RESEARCH》, vol. 33, no. 3, 10 May 2024 (2024-05-10), pages 973 - 994 * |
| 陈进;毛先成;刘占坤;邓浩;: "基于随机森林算法的大尹格庄金矿床三维成矿预测", 大地构造与成矿学, no. 02, 15 April 2020 (2020-04-15) * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112735097A (en) | Regional landslide early warning method and system | |
| CN118378199A (en) | Real-time anomaly detection method in big data analysis platform | |
| CN119312181A (en) | A method, device and electronic device for identifying engine group fault patterns based on density clustering-support vector machine and multi-time prediction weighting | |
| CN110830946A (en) | Mixed type online data anomaly detection method | |
| CN118797321B (en) | Industrial process fault detection method and system based on spatiotemporal feature attention fusion | |
| CN119067225B (en) | Industrial control anomaly interpretation method and system based on difference of generated counterfactual samples | |
| CN119150152A (en) | Mine safety-oriented equipment predictive maintenance monitoring method and electronic equipment | |
| CN117828423A (en) | Photovoltaic module abnormality identification method and system based on statistical characteristics | |
| CN116484289A (en) | A carbon emission abnormal data detection method, terminal and storage medium | |
| CN117877232A (en) | Geological disaster early warning method and system based on GIS | |
| CN115576981A (en) | Anomaly detection method based on combination of supervised algorithm and unsupervised algorithm | |
| CN114154617A (en) | A method and system for identifying abnormal power consumption of low-voltage residential users based on VFL | |
| CN114882293A (en) | Random forest and ship target classification method based on AIS data feature optimization | |
| CN119202829A (en) | A method for verifying abnormal water use data based on deep learning neural network | |
| Ohanuba et al. | Topological data analysis via unsupervised machine learning for recognizing atmospheric river patterns on flood detection | |
| CN118981684A (en) | A method and system for out-of-distribution fault detection based on energy propagation and graph learning | |
| CN119004375A (en) | Construction method of fault diagnosis model of graphic annotation ship engine based on feature fusion | |
| CN120470953A (en) | A gold mine target area optimization method based on spatially constrained modified embedded geochemical anomaly and its system and storage | |
| CN118656640A (en) | Meteorological report generation method and system based on deep learning | |
| CN118094436A (en) | A method and system for detecting anomalies of fault-sensitive self-supervised equipment | |
| Khelil et al. | Sensor Anomaly Detection using Self Features Organizing Maps and Hierarchical-Clustring for Water Quality Assessment | |
| CN119901285B (en) | Visual inertial navigation radar fusion self-positioning method based on self-supervision neural network | |
| CN119336821B (en) | Time series data anomaly detection method and system based on flattened attention mechanism | |
| CN118998639B (en) | A gas leakage detection method based on multi-sensor data of underground pipe network | |
| CN118194100B (en) | Intelligent sampling method and system for blasthole rock powder in explosion zone |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination |