Gold mine target region optimization method and equipment based on element geochemical anomalies

Abstract

The invention is suitable for the technical field of resource exploration, and provides a gold mine target region optimization method and device based on element geochemical anomalies, and the method comprises the steps: obtaining the geochemical sample data of a plurality of sampling points in a research region; training the converter model by using the geochemical sample data to obtain a trained converter model; initializing model parameters of the trained converter model to obtain a student model; performing self-distillation training on the student model by using the geochemical sample data to obtain a geochemical data reconstruction model; inputting the geochemical sample data of the sampling points into a geochemical data reconstruction model for processing to obtain reconstructed geochemical data of the sampling points; and calculating a geochemical anomaly score of each sampling point based on the reconstructed geochemical data of each sampling point, and determining a gold mine target area in the research area based on the geochemical anomaly scores of the plurality of sampling points. The method can meet the requirement of accurately and optimally selecting the target region in gold mine exploration.

Description

Technical Field

[0001] The present application belongs to the field of resource exploration technology, and in particular, relates to a method and equipment for selecting a gold mine target area based on element geochemical anomalies. Background Art

[0002] Using computer technology to process geological data and detect geochemical anomalies is an important means of finding potential gold target areas. However, since geochemical data are usually sparse, complex, and nonlinear, there are many challenges when using computer technology to process them.

[0003] 1. Traditional machine learning methods, such as statistical anomaly detection algorithms (including mean square error, probability graph, etc.) and cluster analysis, factor analysis, etc., are mainly implemented in computers based on simple data statistics and basic mathematical model construction. When processing geochemical data, these methods require a lot of manual feature engineering to adapt to model input requirements. In computer systems, faced with high-dimensional and sparse geochemical data, data reading, calculation and storage operations are complex and time-consuming, the model learning ability is limited, it is easily disturbed by noise, and it is difficult to effectively capture the complex spatial relationships in the data, resulting in low accuracy in the optimization of gold target areas and the inability to accurately identify the geochemical anomaly characteristics and spatial distribution patterns related to gold mineralization.

[0004] 2. Although the detection model based on deep learning has made some progress, there are still many problems. For example, convolutional neural networks (CNN), graph convolutional networks, autoencoders, graph convolutional autoencoders, sparse autoencoders, etc. use their powerful computing power and parallel processing capabilities in computers for automatic feature learning. However, when CNN processes geochemical data, its convolution kernel parameter setting limits the capture of long-distance dependencies, making it difficult to obtain correlation information about gold-related elements in a large spatial range. In the computer implementation, the autoencoder method focuses on minimizing data reconstruction errors, ignoring the spatial relationship and distribution characteristics of geochemical data, and cannot fully utilize spatial information for gold target area optimization.

[0005] 3. The Transformer model architecture performs well in fields such as natural language processing, but it faces many difficulties when applied to geochemical anomaly detection and gold mine target area optimization. In computer systems, its training requires a large amount of data, while geochemical sampling data is relatively sparse, which makes it difficult for model training to converge. In addition, the Transformer model has a complex structure and has extremely high requirements for computing resources (such as central processing units (CPUs), graphics processing units (GPUs)) and memory during the calculation process, such as the large number of matrix multiplication operations when calculating multi-head self-attention. This limits the generalization performance of the trained model in the gold mine target area optimization scenario, making it difficult to accurately distinguish areas that are truly related to gold mines.

[0006] In summary, when the existing technology uses computer technology to optimize gold mine target areas, there are problems such as insufficient utilization of limited geochemical sampling data, difficulty in model training, weak generalization ability, and difficulty in effectively capturing spatial relationships. It cannot meet the needs of accurate optimization of target areas in gold mine exploration. Summary of the invention

[0007] The embodiments of the present application provide a method and device for optimizing gold mine target areas based on element geochemical anomalies, which can solve the problem that traditional gold mine target area optimization methods cannot meet the needs of accurate target area optimization in gold mine exploration.

[0008] In a first aspect, the present application provides a method for selecting a gold target area based on element geochemical anomalies, comprising: Obtain geochemical sample data from multiple sampling points in the study area; geochemical sample data includes concentration values ​​of multiple elements; The converter model is trained using geochemical sample data from multiple sampling points to obtain a trained converter model; the converter model is used to process the geochemical sample data and output reconstructed geochemical data; Initialize the model parameters of the trained converter model to obtain a student model; The student model is trained by self-distillation using geochemical sample data from multiple sampling points to obtain a geochemical data reconstruction model; The geochemical sample data of multiple sampling points are input into the geochemical data reconstruction model for processing to obtain the reconstructed geochemical data of each sampling point; The geochemical anomaly score of each sampling point is calculated based on the reconstructed geochemical data of each sampling point, and the gold target area in the study area is determined based on the geochemical anomaly scores of multiple sampling points.

[0009] Optionally, obtain geochemical sample data from multiple sampling points within the study area, including: For each sampling point, the original geochemical sample data of the sampling point is position-encoded using a position encoding function, and a vector addition operation is performed on the position encoding result and the original geochemical sample data of the sampling point to obtain the geochemical sample data of the sampling point.

[0010] Optionally, the loss function in the process of self-distillation training of the student model using geochemical sample data from multiple sampling points is:

[0011] in, represents the loss function of the j-th generation model obtained at the j-th iteration during the self-distillation training process, represents the loss weight of the jth iteration during the self-distillation training process, Depends on the number of iterations, represents the number of sampling points, Indicates the j-th generation model in the model parameters Next pair The output result is, represents the geochemical sample data of the i-th sampling point, Represents the geochemical sample data of multiple sampling points except the i-th sampling point, Indicates the j-1 generation model in the model parameters Next pair The output result is:

[0012] Optional, loss weight The calculation formula is: , , , represents the initial weight.

[0013] Optionally, the reconstructed geochemical data for each sampling point include reconstructed concentration values ​​of multiple elements; The geochemical anomaly score of each sampling point is calculated based on the reconstructed geochemical data of each sampling point, including: By formula Calculate the geochemical anomaly score of the i-th sampling point ; in, represents the number of elements, represents the concentration value of the kth element at the ith sampling point, Represents the reconstructed concentration value of the kth element at the ith sampling point.

[0014] Optionally, determine gold targets within the study area based on geochemical anomaly scores at multiple sampling points, including: Sort the geochemical anomaly scores of multiple sampling points in descending order; The areas corresponding to the sampling points with the first M geochemical anomaly scores among the sorted multiple geochemical anomaly scores are taken as the gold target areas in the study area.

[0015] Optionally, after the step of performing self-distillation training on the student model using geochemical sample data of multiple sampling points to obtain a geochemical data reconstruction model, the gold mine target area optimization method further includes: The AUC value of the geochemical data reconstruction model is calculated using geochemical sample data of multiple positive samples and multiple negative samples; the positive samples are known gold deposits, and the negative samples are known non-gold deposits; If the AUC value is less than the preset threshold, the model parameters of the geochemical data reconstruction model are adjusted, and the geochemical data reconstruction model after the model parameters are adjusted is used as the student model, and the step of performing self-distillation training on the student model using the geochemical sample data of multiple sampling points to obtain the geochemical data reconstruction model is returned to execute, until the AUC value is greater than or equal to the preset threshold, the step of inputting the geochemical sample data of multiple sampling points into the geochemical data reconstruction model for processing to obtain the reconstructed geochemical data of each sampling point is entered; If the AUC value is greater than or equal to the preset threshold, the step of inputting the geochemical sample data of multiple sampling points into the geochemical data reconstruction model for processing is entered to obtain the reconstructed geochemical data of each sampling point.

[0016] Optionally, the AUC value of the geochemical data reconstruction model is calculated using geochemical sample data of multiple positive samples and multiple negative samples, including: The geochemical data reconstruction model is used to process the geochemical sample data of each positive sample and each negative sample respectively to obtain the reconstructed geochemical data of each positive sample and each negative sample; Obtaining geochemical anomaly scores for each positive sample and each negative sample based on the reconstructed geochemical data of each positive sample and each negative sample; The AUC value of the geochemical data reconstruction model was calculated using the geochemical anomaly scores of multiple positive samples and multiple negative samples.

[0017] Optionally, the AUC value of the geochemical data reconstruction model is calculated as: in, represents the AUC value of the geochemical data reconstruction model, represents the number of positive samples, represents the number of negative samples, represents the indicator function, , represents the geochemical anomaly score of the tth positive sample, represents the geochemical anomaly score of the qth negative sample.

[0018] In a second aspect, an embodiment of the present application provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned method for selecting gold mine target areas based on elemental geochemical anomalies when executing the computer program.

[0019] The above solution of the present application has the following beneficial effects: In an embodiment of the present application, a progressive rebirth mechanism is applied on the basis of a trained converter model to initialize the model parameters of the trained converter model to obtain a student model, thereby prompting the student model to relearn on the basis of retaining some previous knowledge. Since the progressive rebirth mechanism allows accurate identification of geochemical anomalies related to gold mineralization under sparse geochemical data conditions through continuous iterative optimization, accurate gold target area optimization is achieved, overcoming the problem of low accuracy of traditional methods when using limited data for target area prediction. Secondly, in terms of knowledge transfer and model performance improvement, through the progressive rebirth mechanism, the model can gradually accumulate and transfer knowledge related to gold mines, effectively improving the model's ability to understand complex geochemical patterns, thereby improving the accuracy and reliability of gold target area optimization, and meeting the needs of accurate target area optimization in gold mine exploration.

[0020] Other beneficial effects of the present application will be described in detail in the subsequent specific implementation section. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0022] Figure 1 A flow chart of a method for selecting a gold target area based on elemental geochemical anomalies provided in one embodiment of the present application; Figure 2 A curve chart showing the change of the AUC value of each generation of the regeneration model under the regeneration framework in an example of the present application; Figure 3 is a ROC curve diagram of the model in an example of this application; Figure 4 A schematic diagram of the structure of a terminal device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0023] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0024] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.

[0025] It should also be understood that the term “and / or” used in the specification and appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0026] As used in the specification and appended claims of this application, the term "if" can be interpreted as "when" or "uponce" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "uponce it is determined" or "in response to determining" or "uponce [described condition or event] is detected" or "in response to detecting [described condition or event]", depending on the context.

[0027] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0028] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0029] In view of the problem that the traditional gold mine target area optimization method cannot meet the demand for accurate target area optimization in gold mine exploration, the embodiment of the present application provides a gold mine target area optimization method based on element geochemical anomalies. The method initializes the model parameters of the trained converter model by applying a progressive rebirth mechanism on the basis of the trained converter model to obtain a student model, prompting the student model to relearn on the basis of retaining some previous knowledge. Since the progressive rebirth mechanism allows accurate identification of geochemical anomalies related to gold mineralization through continuous iterative optimization under sparse geochemical data conditions, accurate gold mine target area optimization is achieved, overcoming the problem of low accuracy of traditional methods when using limited data for target area prediction. Secondly, in terms of knowledge transfer and model performance improvement, through the progressive rebirth mechanism, the model can gradually accumulate and transfer knowledge related to gold mines, effectively improving the model's ability to understand complex geochemical patterns, thereby improving the accuracy and reliability of gold mine target area optimization, and meeting the demand for accurate target area optimization in gold mine exploration.

[0030] The gold target area optimization method based on elemental geochemical anomalies provided by the present application is exemplarily described below in conjunction with specific embodiments.

[0031] like Figure 1 As shown, the gold target area optimization method based on element geochemical anomalies provided in the embodiment of the present application includes the following steps: Step 11, obtain geochemical sample data from multiple sampling points in the study area.

[0032] The above-mentioned study area is an area where gold target area optimization is required. As a preferred example, the study area can be divided into a grid, and a sampling point can be set in each grid unit to collect the original geochemical sample data of each sampling point. The original geochemical sample data contains the concentration values ​​of multiple elements, such as gold (Au), arsenic (As), antimony (Sb), mercury (Hg) and other elements. The geochemical sample data in the above step 11 is obtained by embedding the position code in the original geochemical sample data. It can be understood that the geochemical sample data, like the original geochemical sample data, also contains the concentration values ​​of multiple elements.

[0033] For example, the geochemical sample data of multiple sampling points can be recorded as a geochemical sample data set , , represents the geochemical sample data of the i-th sampling point, Contains concentration values ​​of various elements. , Indicates the number of sampling points.

[0034] In some embodiments of the present application, the specific implementation method of the above step 11 can be: for each sampling point respectively, use the position encoding function to position encode the original geochemical sample data of the sampling point, perform vector addition operation on the position encoding result and the original geochemical sample data of the sampling point, and obtain the geochemical sample data of the sampling point.

[0035] Specifically, the original geochemical sample data can be preprocessed to meet the input requirements of computer data processing. The data set consisting of multiple element concentration values ​​contained in each sampling point is organized into a format suitable for computer model reading and regarded as an unordered sequence input model. On this basis, position coding is added to assist the model in understanding the spatial information of the data. In the computer data processing process, position coding is of key significance for the model to understand the spatial information of geochemical data. The generation of position coding is based on a computer algorithm, and the relative position coordinates of the geochemical sample are mapped to position coding through a learnable relative position coding function (such as a function based on a multi-layer perceptron (MLP)), and added to the original geochemical sample data. Specifically, the relative position coordinates of each sampling point and the geochemical sample to be reconstructed (i.e., the target sample, each point will take turns to be the target sample) are calculated, mapped through the position coding function, and the position coding is obtained. The coding is added to the original high-dimensional geochemical data (i.e., the original geochemical sample data) to form the final geochemical sample data (i.e., the geochemical sample data in step 11). This process utilizes the powerful computing and data processing capabilities of computers, enhances the model's perception of the spatial relationships of data, and helps to more accurately identify the spatial characteristics of geochemical anomalies related to gold mineralization in subsequent steps.

[0036] The calculation formula of the position encoding function is: .

[0037] in, Represents the need to reconstruct geochemical samples For example, the position coding result of the original geochemical sample data (which can be understood as the original geochemical sample data of the current sampling point of the position coding function) is Represents the original geochemical sample data to be reconstructed The spatial coordinates stored in the computer memory, It represents the spatial coordinates of the original geochemical sample data of the remaining sampling points in the computer system. is a multi-layer perceptron (MLP) function, and its specific calculation process is as follows: When the computer executes this function, first, for the input relative position coordinates , it will be fed into the first linear layer. This linear layer is transformed in the computer through its matrix and bias Perform a linear transformation on the input data, that is, calculate Then, after the nonlinear activation through the ReLU layer, we get Then, through the second linear layer, using its transformation matrix and bias Perform further linear transformation and finally get the function output : These MLP-related parameters ( , ) are trained end-to-end together with the parameters of the converter model in the computer. During the training process, the computer optimizes and adjusts these parameters according to the model's loss function through the back-propagation algorithm. Using this feed-forward network, the computer system maps the relative two-dimensional position coordinates to position codes. In order to maintain compatibility with the original geochemical data in terms of data dimension, the position codes are mapped to L dimensions. In the computer memory, the high-dimensional geochemical data and the generated position codes are vector-added to form a format suitable for model input.

[0038] Through this learnable relative position encoding process, the computer system can obtain a series of advantages. On the one hand, the learnable function enables the computer to adaptively learn the representation of the spatial relationship of the target geochemical sample according to the characteristics of the data in the geochemical anomaly identification task. When the computer processes the data, by constantly adjusting the function The parameters in can better capture the spatial dependencies between samples, thereby improving the model's ability to understand the spatial structure of geochemical data. On the other hand, relative coordinates generate unique position codes for each geochemical sample, which provides a diverse sample set for training the end-to-end transformer model in the computer using a single geochemical dataset. During the training process, the computer can use these different sample sets to better learn data features and improve the generalization ability of the model, so that it can adapt to different geochemical data distributions, thereby showing more accurate and stable performance in geochemical anomaly detection and target area optimization tasks.

[0039] In order to deal with the problem of small data volume, random data masking operation is introduced. Using computer random algorithm, some values ​​in the data are randomly masked. This not only increases the diversity of data and reduces the risk of model overfitting, but also takes advantage of the variable number of elements in the self-attention mechanism, adapts to the requirements of variable-length data input in computer data processing, and prompts the model to learn the potential spatial structure of the data, so as to better discover the abnormal patterns related to gold mines hidden in geochemical data.

[0040] Step 12: training the converter model using geochemical sample data from multiple sampling points to obtain a trained converter model.

[0041] The above transformer model (Transformer) is used to process geochemical sample data and output reconstructed geochemical data. It should be noted that when the transformer model is used to process the geochemical sample data of a certain sampling point (hereinafter referred to as the target geochemical sample data for the convenience of description) and output the reconstructed geochemical data of the sampling point, the geochemical sample data of the sampling point and the geochemical sample data of other sampling points need to be output to the transformer model for processing.

[0042] In some embodiments of the present application, geochemical sample data from multiple sampling points can be used to perform unsupervised training on the transformer model to obtain a trained transformer model. It can be understood that Transformer is a commonly used transformer model, and the model can be trained using a conventional unsupervised training method. That is, the embodiments of the present application can use geochemical sample data from multiple sampling points as training data and use a conventional unsupervised training method of the transformer model to train the model.

[0043] The following is an exemplary explanation of the Transformer training process.

[0044] After the preprocessed data is input into the converter model, the computer system begins to perform a series of complex data processing operations. First, the multi-head self-attention mechanism is used to extract the features of each data point (i.e., sampling point), and the long-range dependencies between data points are explored by calculating the relationship between the query, key, and value. The specific calculation process is to linearly project the input data to obtain features, and then transform it through the scaled dot product attention function to calculate the attention score between each sample and other samples, and then extract features. Next, a fully connected network is used to further extract high-level features. This process involves multiple complex matrix operations and nonlinear transformations, such as residual connections and layer normalization operations, to avoid gradient vanishing or exploding problems and achieve deep feature extraction. The entire process is carried out in the computer's computing unit (such as CPU or GPU), making full use of the computer's computing power to process complex data relationships, thereby mining geochemical characteristic information related to gold mineralization for accurate target area optimization.

[0045] Specifically, the Transformer training process includes the following three parts: (1) Computer extraction of spatial variation characteristics of geochemical samples: The computer receives an unordered sequence of geochemical data, uses a multi-head self-attention mechanism to extract the characteristics of each data point, and captures the long-range dependencies between data points. Then a fully connected network algorithm is used to further extract high-level features. This step enables the computer system to extract the spatial structural characteristics of multivariate geochemical patterns and the compositional relationships between elements.

[0046] Hypothetical Geochemical Sample Dataset , where the geochemical sample data for each data point is Contains the concentration values ​​of various elements. is the target sample (i.e., the target geochemical sample data), and the remaining samples are recorded as After position encoding and data masking, As an unordered sequence input to the converter model. Perform linear projection to obtain features , The interaction between geochemical elements of a single geochemical sample is captured. The preliminary query, key and value are formed. Subsequently, the key, value and query are transformed through a multi-headed scaled dot product attention function to explore the spatial dependencies between given geochemical samples. Head, Geochemical samples The feature is the feature after the self-attention mechanism output conversion : in, yes The query value obtained by linear projection, and yes The dimensions of the keys and values ​​obtained by linear projection are , is any point in the geochemical sample dataset, Will Perform linear projection to obtain features. is the feature set of the geochemical sample. Finally, the feature of each head Perform concatenation and projection to obtain the final features extracted by multi-head attention : In the above formula, is the number of attention mechanism heads, This is an operation to concatenate and project all head features, where the projection uses a fully connected layer. This transformation is recorded as ,Right now: Subsequently, the features extracted by multi-head attention are connected through residual connections and layer normalization to avoid gradient vanishing or gradient exploding.

[0047] In the above formula, is the high-level feature after layer normalization, is the layer normalization function.

[0048] Then continue to use the fully connected network to further extract high-level features, and then go through residual connections and layer normalization: In the above formula, To further extract high-level features, is a fully connected network. The output That is to say, a spatial variation feature extraction is completed. Then After multi-head scaling, click attention function, residual connection, layer normalization, fully connected network, residual connection, layer normalization, feature extraction is performed again. This is repeated N times to form a deep feature extraction process.

[0049] (2) Reconstruct geochemical signals based on extracted features: In the computer system, the interaction between the target sample features and the overall features is achieved through a specific algorithm. Specifically, the attention function between the target sample features and the remaining sample features is calculated, and a linear transformation is applied. After multiple residual connections and layer normalization processes, the interaction between the spatial features of the geochemical sample and the element combination information is completed, and finally the reconstruction result of the target geochemical sample is obtained. This reconstruction process relies on the computer's efficient processing capabilities for large amounts of data and precise algorithm calculations. By comparing the original data and the reconstructed data, it helps to identify geochemical anomaly areas that may be related to gold mines, which serves as the basis for target area optimization.

[0050] The features extracted by the computer system in the previous step are recorded as , is the deep-level feature of the geochemical signal in the overall data set. Then the feature of the target sample is recorded as , the computer system performs the characteristics of the target geochemical sample and the overall characteristics Interaction: in, is the characteristic of the target sample The characteristics of the remaining samples The attention function between Represents the linear transformation that the computer system applies to the remaining sample features. The computer system uses this transformation to capture the spatial dependence of the geochemical sample for the target sample and extract complex features to reconstruct the geochemical background of the target sample.

[0051] The result of interaction with the overall features , the computer system performs residual connections and layer normalization to avoid gradient disappearance or gradient explosion in the model: Then After the transformation of function M, residual connection and layer normalization again, we get :

[0052] In the above formula, function M is the multi-head attention function.

[0053] The computer system performs the conversion of fully connected network, residual connection and layer normalization again to obtain : In the above formula, A fully connected network.

[0054] Therefore, from arrive , the interaction between the spatial characteristics of the geochemical sample and the element combination information is completed in the computer. This process is also similar to the feature extraction stage, and the computer system will repeat it N times. Finally, the final output result is obtained - the reconstructed geochemical data of the target geochemical sample data.

[0055] The computer system will perform the above operations for each sample, interact with the overall data, and generate a reconstructed geochemical signal. This completes the reconstruction of all geochemical samples. The reconstruction results reflect the characteristics of the geochemical background, provide a benchmark for subsequent anomaly detection, and become favorable information for the selection of gold mine targets.

[0056] (3) With the goal of minimizing the difference between the original data and the reconstructed data, the L1 loss is used to quantify the difference. During the training process of the computer system, the objective function value (i.e., the loss value) is continuously calculated with the help of computer optimization algorithms (such as the Adam optimizer) and the back propagation algorithm, and the model parameters are adjusted according to the calculation results. The computer gradually updates the model through rapid processing and iterative calculation of a large amount of data, so that the model has basic geochemical background recognition capabilities, laying the foundation for subsequent rebirth framework training, so as to more accurately screen out abnormal information related to gold mine targets from geochemical data.

[0057] The initial model (i.e., the trained transformer model) is represented as : .in Represents a learnable parameter. The training objective function (i.e., loss function) of the converter model is: in, is the target geochemical sample data, For the converter model in the model parameters Next pair The output result is, is the loss function of the converter model, and the remaining samples are recorded as , The converter model.

[0058] Step 13, initializing the model parameters of the trained converter model to obtain a student model.

[0059] In some embodiments of the present application, after the converter model training is completed, the computer system performs a parameter reset operation to generate a first-generation reborn model (i.e., the student model mentioned above, which can also be called a newborn model). This process involves reinitializing the model parameters, prompting the model to relearn on the basis of retaining some previous knowledge to improve the generalization ability of the model.

[0060] Step 14, using the geochemical sample data of multiple sampling points to perform self-distillation training on the student model to obtain a geochemical data reconstruction model.

[0061] In some embodiments of the present application, a self-knowledge distillation mechanism (i.e., self-distillation training) is introduced, and the first-generation reborn model is regarded as a "teacher model", and its output is used as a guiding signal for the training of the new model. During the computer processing process, the new model acquires richer knowledge of geochemical background characteristics by learning the output of the "teacher model", thereby enhancing the learning ability of complex geochemical patterns, especially helping to identify complex characteristic patterns related to gold mineralization, and providing a more accurate basis for the optimization of gold target areas.

[0062] Specifically, the first generation model After the reconstruction training of the computer system, the parameters are reset and the model is reset to the first generation reborn model . First generation model The reconstruction results of geochemical data are expressed as , It can be regarded as the initial geochemical background signal extracted by the model. This background needs further training and cannot be directly used for gold target area selection. Therefore, the computer system training model When The output of the is used together with the original geochemical data to guide the training of the model. The output results will be considered to be closer to the real geochemical background signal and more conducive to the selection of gold target areas.

[0063] Understandably, self-distillation training requires retraining the first generation model Multiple iterations are performed to obtain a geochemical data reconstruction model that meets performance requirements.

[0064] The loss function in the process of self-distillation training of the student model using geochemical sample data from multiple sampling points is: in, represents the loss function of the j-th generation model obtained at the j-th iteration during the self-distillation training process, Represents the reconstruction loss , Reincarnation loss , represents the loss weight of the jth iteration during the self-distillation training process, Depends on the number of iterations, represents the number of sampling points, Indicates the j-th generation model in the model parameters Next pair The output result is, represents the geochemical sample data of the i-th sampling point, Represents the geochemical sample data of multiple sampling points except the i-th sampling point, Indicates the j-1 generation model in the model parameters Next pair The output of the model. is the model parameter of the j-th generation model obtained in the j-th iteration, model parameter The model parameters of the j-1 generation model obtained in the j-1 iteration. j is a positive integer greater than or equal to 1 (for the convenience of description, the j-generation model obtained in the j-1 iteration is called the j-generation rebirth model ), The first generation reborn model The parameters learned on the computer are obtained by resetting the initial model Parameters Then re-learn the model. and They have the same model structure, but the parameters learned on the computer are different. It is understandable that for It should be noted that in the model After that, the parameters will be reset again to generate a new model , which is trained by The output of the model and the geochemical sample data are used as guidance. In the computer system, this iterative process is repeated many times (the training termination condition can be set to the number of iterations reaching a preset number, or the loss value of the model). meet the preset loss requirement value), and finally form a continuously regenerated network framework.

[0065] During the training of the new model, the computer uses the output of the previous model as an additional reference signal, so that the new model gradually approaches the output distribution of the previous model while learning the original geochemical data. The reconstruction loss is controlled by adjusting the rebirth coefficient (i.e., the loss weight mentioned above, such as by adjusting it in a linear increase manner). and rebirth loss The computer calculates and adjusts the loss weight in each iteration according to the set algorithm and coefficient value to ensure the stable convergence of the model and continuously improve the reconstruction accuracy, so as to more accurately locate the possible gold mine target area. Therefore, the training goal of the new model is to minimize the reconstruction loss at the same time. and rebirth loss .

[0066] Specifically, the above loss weight The calculation formula is: , , , Represents the initial weight, which can be set according to the actual situation. As the number of iterations (i.e., the number of rebirths) increases, the value of the loss weight increases linearly. This dynamic adjustment strategy ensures that in the later stages of training, as the model's understanding of the data deepens, it can rely more on the knowledge output by the previous generation model, thereby improving the accuracy and convergence speed of the model.

[0067] It can be seen that during the self-distillation training process, the computer repeatedly executes the training process of the rebirth model and performs multiple iterative optimizations. The training of each generation of models is guided by the output of the previous generation of models and the geochemical sample data, forming a continuously regenerated network framework in the computer system. During the iteration process, the computer increases the rebirth loss weight generation by generation (according to the set linear increase rule) and reduces the learning rate according to the model convergence characteristics. Through repeated processing of a large amount of data and parameter adjustment, the computer helps the model gradually approach the global optimal solution, effectively reduces the risk of overfitting, and enables the model to continuously optimize the judgment of geochemical anomaly areas during the optimization of gold mine target areas, thereby improving the accuracy and reliability of optimization.

[0068] In the gold target area optimization regeneration network framework, this means that as training progresses, the guidance of the teacher model (previous generation model) to the new model (current generation model) gradually increases. From the perspective of computer data processing, when calculating the loss function, the gradual increase in the distillation loss coefficient (i.e., loss weight) makes the model pay more attention to learning the features and information extracted by the previous generation model, so as to find the optimal solution faster in the data space and improve the convergence speed of the model. In the computer's memory, the model can better focus on the geochemical characteristic information related to gold mineralization through the effective use of the previous generation of output knowledge, thereby improving the accuracy and reliability of gold target area selection.

[0069] The learning rate is reduced generation by generation during the training process: During the model training process, as the number of generations increases, the setting value of the learning rate in the computer system is reduced generation by generation. This strategy is to enable subsequent models to be trained at a lower learning rate after reaching the performance of the previous generation model, thereby helping the model to get closer to the global optimal solution during the computer optimization process and effectively reducing the risk of overfitting.

[0070] Since the subsequent network models inherit the knowledge stored in the computer memory of the previous generation network models, they converge faster during training than the first generation network. In the computer program implementation, when it is detected that the model performance reaches or approaches the previous generation model, the learning rate reduction mechanism is triggered. There are two main reasons for this adjustment: First, in order to adapt to the rapid convergence characteristics of the model learning process. In the early stages of computer training, a larger learning rate helps the model to quickly explore the data space, but as the training progresses, the model gradually approaches the optimal solution area, and an excessively large learning rate may cause the model to skip the optimal solution. Therefore, reducing the learning rate generation by generation allows the model to more finely adjust the model parameters while accelerating the convergence speed, avoiding missing the global optimal solution. In geochemical data processing, this means that the model can more accurately learn the complex features and relationships contained in the data, such as the subtle correlations between element concentrations and the spatial distribution patterns of elements related to gold mineralization.

[0071] Second, the model can continue to learn at a lower learning rate after reaching the performance level of the previous generation. A lower learning rate reduces the update amplitude of model parameters in the computer, making it more sensitive to detailed information in the data. When processing geochemical data, this fine-tuning capability helps the model capture more subtle abnormal features, enhances its ability to learn the complex knowledge contained in the data, and further improves the accuracy and robustness of the model in anomaly detection and gold mine target area optimization tasks. In the computer implementation process, specific rules and algorithms are set to control the generation-by-generation reduction of the learning rate. For example, exponential decay or linear decay can be used. According to different decay strategies and parameter settings, the model can better balance the convergence speed and model performance during training to achieve the best target area selection effect.

[0072] It should be noted that in order to ensure that the model can meet the needs of accurate selection of target areas in gold mine exploration, it is also necessary to calculate the correlation between the model reconstruction results and the distribution of known gold mine points. Specifically, after the step of self-distillation training the student model using the geochemical sample data of multiple sampling points to obtain the geochemical data reconstruction model, the above-mentioned gold mine target area optimization method also includes: using the geochemical sample data of multiple positive samples and multiple negative samples to calculate the AUC value of the geochemical data reconstruction model, if the AUC value is greater than or equal to the preset threshold, then entering step 15, inputting the geochemical sample data of multiple sampling points into the geochemical data reconstruction model for processing, and obtaining the reconstructed geochemical data of each sampling point; if the AUC value is less than the preset threshold, then adjusting the model parameters of the geochemical data reconstruction model, and using the geochemical data reconstruction model after the model parameters are adjusted as the student model in step 14, returning to execute step 14, using the geochemical sample data of multiple sampling points to self-distill the student model to obtain the geochemical data reconstruction model, until the AUC value is greater than or equal to the preset threshold, then entering step 15, inputting the geochemical sample data of multiple sampling points into the geochemical data reconstruction model for processing, and obtaining the reconstructed geochemical data of each sampling point. Wherein, the positive sample is a known gold mine point, and the negative sample is a known non-gold mine point.

[0073] The calculation process of the AUC value of the geochemical data reconstruction model is as follows: first, the geochemical sample data of each positive sample and each negative sample are processed respectively using the geochemical data reconstruction model to obtain the reconstructed geochemical data of each positive sample and each negative sample (that is, the geochemical sample data of the positive sample is input into the geochemical data reconstruction model for processing to obtain the reconstructed geochemical data of the positive sample; the geochemical sample data of the negative sample is input into the geochemical data reconstruction model for processing to obtain the reconstructed geochemical data of the negative sample); then, the geochemical anomaly score of each positive sample and each negative sample is obtained based on the reconstructed geochemical data of each positive sample and each negative sample; finally, the AUC value of the geochemical data reconstruction model is calculated using the geochemical anomaly scores of multiple positive samples and multiple negative samples.

[0074] The calculation formula for the AUC value of the geochemical data reconstruction model is: in, represents the AUC value of the geochemical data reconstruction model, represents the number of positive samples, represents the number of negative samples, represents the indicator function, , represents the geochemical anomaly score of the tth positive sample, represents the geochemical anomaly score of the qth negative sample.

[0075] The calculation formula is: Calculate the geochemical anomaly score of the tth positive sample .in, represents the number of elements, represents the concentration value of the kth element of the tth positive sample, Represents the reconstructed concentration value of the kth element of the tth positive sample.

[0076] The calculation formula is: Calculate the geochemical anomaly score of the qth negative sample .in, represents the number of elements, represents the concentration value of the kth element of the qth negative sample, Represents the reconstructed concentration value of the kth element of the qth negative sample.

[0077] It should be noted that the AUC value is a quantitative indicator of the spatial correlation between the anomaly score and the known gold mine distribution. The false positive rate is compared with the true positive rate to evaluate the accuracy and robustness of the model in geochemical anomaly detection. The computer calculates the anomaly scores of positive and negative samples according to the set algorithm, and then obtains the AUC value. The higher the AUC value, the better the model detection result, which can accurately evaluate the model performance.

[0078] Step 15, inputting the geochemical sample data of multiple sampling points into the geochemical data reconstruction model for processing to obtain the reconstructed geochemical data of each sampling point.

[0079] After successfully completing the model training and optimization process, the key stage of gold mine target area optimization begins. First, the input geochemical sample data is processed using the final trained geochemical data reconstruction model to obtain the reconstructed geochemical data of each sampling point. The reconstructed geochemical data of each sampling point includes the reconstructed concentration values ​​of multiple elements.

[0080] Step 16, calculating the geochemical anomaly score of each sampling point based on the reconstructed geochemical data of each sampling point, and determining the gold target area in the study area based on the geochemical anomaly scores of multiple sampling points.

[0081] In the computer system, for the geochemical sample data of each sampling point, the difference between its original element concentration value and the element concentration value reconstructed by the model is accurately calculated to determine the geochemical anomaly score. Specifically, the geochemical anomaly score of each sampling point is calculated based on the reconstructed geochemical data of each sampling point as follows: By formula Calculate the geochemical anomaly score of the i-th sampling point ; In the formula, represents the number of elements, represents the concentration value of the kth element at the ith sampling point, Represents the reconstructed concentration value of the kth element at the ith sampling point.

[0082] After calculating the geochemical anomaly scores of each sampling point, the geochemical anomaly scores of multiple sampling points can be sorted in descending order, and the areas corresponding to the sampling points to which the first M geochemical anomaly scores belong among the sorted multiple geochemical anomaly scores are taken as gold target areas in the study area. These anomaly areas may contain potential information of gold mineralization. For example, the value of M can be 5.

[0083] In some embodiments of the present application, based on the selected gold target area, geochemical data and other relevant information of the area and its surroundings are further collected, added to the original data set, and the data preprocessing stage is entered again to restart the entire model training and target area optimization process. In the computer system, this process realizes the iterative optimization of the gold target area. With the continuous addition of new data and repeated training of the model, the model can more accurately identify and select more potential gold target areas, and gradually improve the accuracy and reliability of the gold target area optimization.

[0084] In practical applications, we can fully integrate relevant information on geology, topography, regional structure, etc. on the basis of the identified geochemical anomaly areas (i.e., the gold target areas mentioned above), and combine the known gold mine distribution information and fault zone distribution to carry out further screening and evaluation. In the computer system, a special gold mine target area optimization algorithm is preset, which comprehensively considers the following key factors to calculate the probability score of each anomaly area becoming a gold mine target area: Element combination characteristics: Analyze based on specific element combination patterns that are closely related to known gold mineralization models. For example, if the enrichment or depletion ratio of certain specific elements is consistent with the element symbiotic combination law in the typical gold mineralization process, the corresponding weight will be increased when calculating the score. For example, gold (Au) and arsenic (As), antimony (Sb), mercury (Hg) and other elements often have a specific symbiotic relationship in some gold mineralization systems. When the combination of these elements in the abnormal area shows characteristics that conform to the mineralization law, the possibility score of the area becoming a gold target area will increase.

[0085] Anomaly intensity: Anomaly intensity reflects the significance of geochemical anomalies and is an important indicator for determining the potential of gold target areas. In the computer, the anomaly intensity is quantified by calculating the degree to which the element concentration in the anomaly area deviates from the background value. The greater the anomaly intensity, the more intense the mineralization may be in the area, and a higher score will be assigned when calculating the possibility score. For areas where the gold concentration is much higher than the background value and has a large concentration gradient change, their higher potential value will be reflected in the score calculation.

[0086] Spatial distribution pattern: Study the spatial distribution pattern and regularity of abnormal areas. If the abnormal area shows spatial distribution characteristics related to known gold mineralization structures or geological bodies, such as distribution along fault zones, fold axes or specific stratigraphic interfaces, it will be given priority consideration when calculating the score. For example, abnormal areas with linear or zonal distribution and consistent with the trend of regional fault zones may imply the association between ore fluid migration channels and favorable mineralization sites, thereby increasing the possibility score of becoming a gold target area.

[0087] Regional structural information: Regional structure plays an important role in controlling the formation and distribution of gold deposits. In computer analysis, the structural position, structural type and structural complexity of the abnormal area are examined in detail. For example, abnormal areas located in favorable structural locations such as the intersection of large faults, the core of anticlines or ductile shear zones will receive a higher weight when calculating the possibility score because they are more conducive to the aggregation and precipitation of mineralizing materials.

[0088] In addition, considering that the area within 5km around the gold deposits and fault zones may have been fully explored or there may be geological conditions that are unfavorable for mineralization (such as the dispersion of mineralized elements or the destruction of geological structures), when calculating the possibility score, the abnormal areas within this range are downgraded to reduce the possibility of being selected as gold deposit targets.

[0089] By incorporating all of the above factors into the gold target area optimization algorithm, the computer system calculates a comprehensive and scientifically reasonable possibility score for each abnormal area. Finally, all abnormal areas are sorted according to the calculated possibility score. In the computer system, through an efficient sorting algorithm, the scores of different areas are quickly compared to determine their priority order. Areas with higher scores are considered to have higher gold mineralization potential and are thus selected as the most likely gold targets. Finally, the computer system outputs these optimization results and presents them to geological prospectors in an intuitive and clear manner, providing valuable decision-making references for gold exploration, helping to improve the efficiency and accuracy of gold exploration, and reducing exploration costs and risks.

[0090] The gold mine target area optimization method of the present application is exemplified below with reference to specific examples.

[0091] Taking the gold mining area in Jiaoxibei, Shandong Province as an example, the geochemical measurements of multiple elements of stream sediments in the area were selected as raw data. In a computer environment (such as the pytorch framework (the pytorch framework is an open source deep learning framework for machine learning and deep learning), the graphics card configuration is Nvidia RTX 3090Ti 128G), the converter network of the progressive rebirth framework was used for geochemical anomaly detection and gold target area optimization. First, the raw data was preprocessed, including data format adjustment, position encoding addition, and random data masking. Then the processed data was input into the model for training. The model adopts a specific layer structure and a multi-head attention mechanism (8-layer structure and 2-head attention mechanism).

[0092] During the training process, the computer records the relevant indicators of each generation of models, such as the receiver operating characteristic (ROC) curve and area under the curve (AUC) value of the geochemical anomaly identification results, the gold target area optimization score, etc. Figure 2As shown in the figure, with the increase in the number of model regenerations, the AUC shows an overall improvement trend. This means that the accuracy of geochemical anomaly identification has gradually improved from a certain level of the initial model, and the gold target area optimization score has also gradually improved. Figure 3 As shown in Table 1, after multiple iterations, the selected gold target area has a high correlation with the known mineral point distribution, which preliminarily proves the feasibility and effectiveness of the present application in the optimization of gold target areas. The entire experimental process fully relies on the powerful computing power and data processing capabilities of the computer. From data acquisition, preprocessing, model training to target area optimization and iterative optimization, all are completed in a large computer workstation with a graphics processing unit (GPU). By utilizing the parallel processing capabilities of the graphics processing unit, the scientific calculation and data processing of the model are performed, so that the model optimized by the rebirth framework is significantly improved in terms of the accuracy and reliability of target area optimization. It demonstrates the great potential of the present invention in actual gold mine exploration applications. Figure 3 In the figure, the horizontal axis is the false positive rate, the vertical axis is the true positive rate, and ROC Curve represents the ROC curve. Table 1 only shows the locations of sampling points corresponding to some gold mine target areas. The x-axis coordinate and the y-axis coordinate are the horizontal and vertical coordinates of the sampling points in the plane rectangular coordinate system.

[0093] Table 1 Gold target area optimization scores

[0094] In summary, the present application has significant advantages over the prior art: in terms of gold target area optimization, the progressive rebirth framework of the present application allows for more accurate identification of geochemical anomalies related to gold mineralization under sparse geochemical data conditions through continuous iterative optimization, thereby achieving more accurate gold target area optimization, overcoming the low accuracy of traditional methods when using limited data for target area prediction. Secondly, in terms of knowledge transfer and model performance improvement, through the progressive rebirth mechanism, the model can gradually accumulate and transfer knowledge related to gold mines, effectively improving the model's ability to understand complex geochemical patterns, thereby improving the accuracy and reliability of gold target area optimization. Finally, the present application has good scalability and adaptability. It is not only suitable for gold target area optimization, but can also be adjusted according to different mineralizing element combinations and geological conditions, and applied to target area optimization of other metal ores or geological targets, providing a general and effective data processing and target area optimization method for the field of geological exploration.

[0095] The terminal device provided in this application is exemplarily described below in conjunction with specific embodiments.

[0096] like Figure 4 As shown, an embodiment of the present application provides a terminal device, such as Figure 4As shown, the terminal device D10 of this embodiment includes: at least one processor D100 ( Figure 4 Only one processor is shown in the figure), a memory D101, and a computer program D102 stored in the memory D101 and executable on the at least one processor D100, wherein the processor D100 implements the steps of any of the above-mentioned method embodiments when executing the computer program D102.

[0097] Specifically, when the processor D100 executes the computer program D102, the model parameters of the trained converter model are initialized to obtain a student model by applying a progressive rebirth mechanism based on the trained converter model, so as to prompt the student model to relearn on the basis of retaining some previous knowledge. Since the progressive rebirth mechanism allows accurate identification of geochemical anomalies related to gold mineralization through continuous iterative optimization under sparse geochemical data conditions, accurate gold target area optimization is achieved, overcoming the problem of low accuracy of traditional methods when predicting targets using limited data. Secondly, in terms of knowledge transfer and model performance improvement, through the progressive rebirth mechanism, the model can gradually accumulate and transfer knowledge related to gold mines, effectively improving the model's ability to understand complex geochemical patterns, thereby improving the accuracy and reliability of gold target area optimization, and meeting the needs of accurate target area optimization in gold mine exploration.

[0098] The processor D100 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0099] In some embodiments, the memory D101 may be an internal storage unit of the terminal device D10, such as a hard disk or memory of the terminal device D10. In other embodiments, the memory D101 may also be an external storage device of the terminal device D10, such as a plug-in hard disk, a smart memory card (SMC, SmartMedia Card), a secure digital (SD, Secure Digital) card, a flash card (Flash Card), etc. equipped on the terminal device D10. Further, the memory D101 may also include both an internal storage unit of the terminal device D10 and an external storage device. The memory D101 is used to store an operating system, an application program, a boot loader (BootLoader), data, and other programs, such as the program code of the computer program, etc. The memory D101 may also be used to temporarily store data that has been output or is to be output.

[0100] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0101] An embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the above-mentioned method embodiments can be implemented.

[0102] An embodiment of the present application provides a computer program product. When the computer program product runs on a terminal device, the terminal device can implement the steps in the above-mentioned method embodiments when executing the computer program product.

[0103] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, which can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may at least include: any entity or device that can carry the computer program code to the terminal device, a recording medium, a computer memory, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), an electric carrier signal, a telecommunication signal, and a software distribution medium. For example, a USB flash drive, a mobile hard disk, a magnetic disk or an optical disk. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electric carrier signals and telecommunication signals.

[0104] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0105] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0106] In the embodiments provided in the present application, it should be understood that the disclosed methods can be implemented in other ways. For example, the method / device embodiments described above are merely schematic. For example, the division of the modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0107] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0108] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for selecting a gold target area based on elemental geochemical anomalies, characterized in that: include: Obtain geochemical sample data from multiple sampling points in the study area; The geochemical sample data includes concentration values ​​of multiple elements; The converter model is trained using the geochemical sample data of the plurality of sampling points to obtain a trained converter model; the converter model is used to process the geochemical sample data and output reconstructed geochemical data; Initialize the model parameters of the trained converter model to obtain a student model; Using the geochemical sample data of the plurality of sampling points to perform self-distillation training on the student model to obtain a geochemical data reconstruction model; Inputting the geochemical sample data of the plurality of sampling points into the geochemical data reconstruction model for processing to obtain reconstructed geochemical data of each sampling point; The geochemical anomaly score of each sampling point is calculated based on the reconstructed geochemical data of each sampling point, and the gold target area in the study area is determined based on the geochemical anomaly scores of the multiple sampling points.

2. The gold mine target area optimization method according to claim 1, characterized in that: The acquisition of geochemical sample data at multiple sampling points within the study area includes: For each sampling point, the original geochemical sample data of the sampling point is position-encoded using a position encoding function, and a vector addition operation is performed on the position encoding result and the original geochemical sample data of the sampling point to obtain the geochemical sample data of the sampling point.

3. The gold mine target area optimization method according to claim 1, characterized in that: The loss function in the process of self-distillation training of the student model using the geochemical sample data of the multiple sampling points is: ； in, represents the loss function of the j-th generation model obtained at the j-th iteration during the self-distillation training process, represents the loss weight of the jth iteration during the self-distillation training process, Depends on the number of iterations, represents the number of sampling points, Indicates the j-th generation model in the model parameters Next pair The output result is, represents the geochemical sample data of the i-th sampling point, Represents the geochemical sample data of multiple sampling points except the i-th sampling point. Indicates the j-1 generation model in the model parameters Next pair The output result is:

4. The gold mine target area optimization method according to claim 3, characterized in that: The loss weight The calculation formula is: , , , represents the initial weight.

5. The gold mine target area optimization method according to claim 1, characterized in that: The reconstructed geochemical data for each sampling point include reconstructed concentration values ​​of multiple elements; The geochemical anomaly score of each sampling point is calculated based on the reconstructed geochemical data of each sampling point, including: By formula Calculate the geochemical anomaly score of the i-th sampling point ; in, represents the number of elements, represents the concentration value of the kth element at the ith sampling point, Represents the reconstructed concentration value of the kth element at the ith sampling point.

6. The gold mine target area optimization method according to claim 5, characterized in that: Determining the gold target area in the study area based on the geochemical anomaly scores of the multiple sampling points includes: sorting the geochemical anomaly scores of the plurality of sampling points in descending order; The areas corresponding to the sampling points to which the first M geochemical anomaly scores belong among the sorted multiple geochemical anomaly scores are taken as the gold mine target areas within the study area.

7. The gold mine target area optimization method according to claim 1, characterized in that: After the step of using the geochemical sample data of the plurality of sampling points to perform self-distillation training on the student model to obtain a geochemical data reconstruction model, the gold mine target area optimization method further includes: The AUC value of the geochemical data reconstruction model is calculated using geochemical sample data of multiple positive samples and multiple negative samples; the positive samples are known gold deposits, and the negative samples are known non-gold deposits; If the AUC value is less than a preset threshold, the model parameters of the geochemical data reconstruction model are adjusted, and the geochemical data reconstruction model after the model parameters are adjusted is used as a student model, and the step of performing self-distillation training on the student model using the geochemical sample data of the multiple sampling points to obtain the geochemical data reconstruction model is returned to execute, until the AUC value is greater than or equal to the preset threshold, the step of inputting the geochemical sample data of the multiple sampling points into the geochemical data reconstruction model for processing to obtain the reconstructed geochemical data of each sampling point is entered; If the AUC value is greater than or equal to the preset threshold, the step of inputting the geochemical sample data of the multiple sampling points into the geochemical data reconstruction model for processing to obtain reconstructed geochemical data for each sampling point is entered.

8. The gold mine target area optimization method according to claim 7, characterized in that: The method of calculating the AUC value of the geochemical data reconstruction model using geochemical sample data of multiple positive samples and multiple negative samples includes: Using the geochemical data reconstruction model to process the geochemical sample data of each positive sample and each negative sample respectively, to obtain reconstructed geochemical data of each positive sample and each negative sample; Obtaining geochemical anomaly scores for each positive sample and each negative sample based on the reconstructed geochemical data of each positive sample and each negative sample; The AUC value of the geochemical data reconstruction model is calculated using the geochemical anomaly scores of the multiple positive samples and the multiple negative samples.

9. The gold mine target area selection method according to claim 8, characterized in that: The calculation formula of the AUC value of the geochemical data reconstruction model is: in, represents the AUC value of the geochemical data reconstruction model, represents the number of positive samples, represents the number of negative samples, represents the indicator function, , represents the geochemical anomaly score of the tth positive sample, represents the geochemical anomaly score of the qth negative sample.

10. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method for selecting gold target areas based on element geochemical anomalies as described in any one of claims 1 to 9 is implemented.

Images