Intelligent cloud management and analysis system for geological exploration data

Abstract

The invention discloses a geological exploration data intelligent cloud management and analysis system. The system comprises a data access module which is responsible for supporting automatic or manual importing of geological exploration data from multiple sources; the data management and integration module is used for receiving the data transmitted from the data acquisition and uploading module, performing version control and integrating the data; the data processing module receives the data provided by the data management and integration module, is responsible for data cleaning, format conversion and normalization processing, introduces an adversarial training technology and generates an adversarial sample; the intelligent analysis module is used for constructing a prediction model, the prediction model is used for modeling time series geological data by using LSTM, capturing a long-term dependency relationship in the data and identifying a geological change trend, then the output of the LSTM is used as a feature to be input into an SVM classifier, classification or regression prediction is carried out on a geological structure, and the geological process and a disaster occurrence mechanism are simulated to obtain a geological model; and generating more synthetic data similar to real data and used for training a prediction model.

Description

Technical Field

[0001] The present invention relates to the field of geological exploration technology, and in particular to an intelligent cloud management and analysis system for geological exploration data. Background Art

[0002] As a vital supporter of mineral resource development, environmental protection, and disaster early warning, the geological exploration industry leverages artificial intelligence to enable rapid storage, efficient processing, intelligent analysis, and visualization of geological exploration data, providing more convenient, accurate, and comprehensive support. Data is a crucial basis for decision-making in geological exploration. Intelligent cloud management and analysis technology enables comprehensive, accurate, and timely access to geological exploration data, providing decision-makers with a more scientific and rational basis for decision-making.

[0003] In existing technology, suppose a geological survey team is using a machine learning model to predict the risk of geological hazards in a certain region. However, due to very limited historical geological hazard data in the region, the team can only collect a small number of training samples. In this case, even the most ingeniously designed model will not be able to fully learn the patterns and characteristics of geological hazards due to insufficient training data. As a result, when the model is applied to actual predictions, its accuracy may be significantly compromised, resulting in significant deviations between the predicted results and the actual situation. Therefore, a smart cloud management and analysis system for geological survey data is proposed. Summary of the Invention

[0004] The purpose of the present invention is to solve the shortcomings of the prior art, such as the decline in prediction accuracy caused by insufficient training data, and to propose a geological exploration data intelligent cloud management and analysis system.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] A geological exploration data intelligent cloud management and analysis system, including:

[0007] Data access module: responsible for supporting automatic or manual import of geological survey data from various sources, including field survey data and sensor data;

[0008] Data management and integration module: Receives data from the data collection and upload module, performs version control, ensures data consistency and traceability, integrates data across projects and domains, and forms a unified data view;

[0009] Data processing module: accepts data provided by the data management and integration module, and is responsible for data cleaning, format conversion, and normalization to ensure data quality. It also introduces adversarial training technology to generate adversarial samples, thereby improving the model's robustness to noise and bias.

[0010] Intelligent Analysis Module: This module uses the finite element method to establish a mathematical model simulating the mechanism of disaster occurrence. By setting different parameter combinations and initial conditions, the mathematical model is run to generate synthetic data that is similar to real data but more abundant. This data is then used to train the prediction model. Simultaneously, the module accesses the dataset provided by the data processing module to train the prediction model and build a prediction model. The prediction model uses LSTM to model time series geological data, capture long-term dependencies in the data, and identify geological change trends. The LSTM output is then input as a feature into the SVM classifier to perform classification or regression prediction of geological structures.

[0011] Report generation and sharing module: Receives the report content after analysis from the intelligent analysis module, generates an exploration report, and supports online preview, download, and sharing of the report under permission control;

[0012] Data security module: responsible for data security and privacy protection during data storage, analysis and processing.

[0013] The above technical solution further includes:

[0014] Furthermore, the data access module includes a data import unit, a data verification unit, a data upload unit and a log recording unit. The data import unit is responsible for receiving data from various sources and converting it into a format recognizable by the system. The data import unit includes a mobile device APP interface, a web page interface and a sensor interface. The mobile device APP interface interacts with the mobile device APP for data to receive field survey data. The web page interface interacts with the web page for data to receive data uploaded by users. The sensor interface interacts with various sensors (such as GPS, seismic wave sensors, magnetic sensors, etc.) for data to receive real-time or periodically collected data. The data verification unit verifies the imported data to ensure the quality of the data. The data verification unit includes a format verification subunit, an integrity verification subunit and a rationality verification subunit. The format verification subunit checks whether the data format meets the system requirements. The integrity verification subunit checks whether the data is complete, such as whether required fields are missing. The rationality verification subunit checks the rationality of the data based on the common sense and rules of geological exploration, such as whether the GPS coordinates are within a reasonable range, whether the geological description conforms to geological common sense, etc. The data upload unit uploads the verified data In the system, the data upload unit includes a database interface and a cache interface. The database interface interacts with the system's database and stores data in a corresponding table. The cache interface stores data in a cache (such as Redis) to improve data access speed. The log recording unit records the process and results of data import, verification, and upload. The log recording unit includes an operation log subunit and an error log subunit. The operation log subunit records information such as the time, operator, and operation result of each data import, verification, and upload operation. The error log subunit records error information and processing results encountered during the operation. The data import unit passes the received data to the data verification unit for verification. The data verification unit returns the verification result (pass / fail) to the data import unit. The data verification unit passes the verified data to the data upload unit for upload. The data upload unit returns the upload result (success / failure) to the data verification unit (or log recording unit for recording). The data upload unit uploads the verified data to a database or cache for storage. The database or cache returns the storage result (success / failure) to the data upload unit (or log recording unit for recording).

[0015] Furthermore, the data processing module includes a data receiving unit, a data cleaning unit, a format conversion unit, a normalization processing unit and an adversarial training unit. The data receiving unit is responsible for receiving data from the data management and integration module and performing preliminary data verification. The data cleaning unit performs cleaning operations on the received data, including removing duplicate data, filling missing values, correcting erroneous data, etc. The format conversion unit converts the data into an appropriate format as needed for subsequent analysis or model training. The normalization processing unit normalizes the data to ensure that the data is compared and analyzed on a unified scale. The adversarial training unit is responsible for generating adversarial samples.

[0016] Furthermore, the adversarial training unit generates adversarial samples in the following specific steps:

[0017] Compute the gradient:

[0018] For a given input sample x and label y, calculate the gradient of the loss function L(x,y) of the LSTM and SVM classifiers on x with respect to x. The gradient represents the rate of change of the loss function in the input space and is the key to guiding how to adjust the input to maximize the loss function.

[0019] Determine the perturbation direction and magnitude:

[0020] According to the calculated gradient, the direction of the disturbance is determined, and the direction of the disturbance is selected as the sign direction of the gradient, that is, The amplitude of the perturbation is controlled by the hyperparameter ε, which determines the amount of noise added to the input sample. The selection of ε requires a balance between the attack effect and the identifiability of the sample. An ε that is too large causes the sample to become unidentifiable, while an ε that is too small cannot effectively mislead the model.

[0021] Generate adversarial examples:

[0022] According to the determined perturbation direction and amplitude, the adversarial sample x' is generated, and x' is calculated by the following formula:

[0023] Furthermore, the intelligent analysis module includes a synthetic data generation unit, a model training and optimization unit, and a data processing module interface unit. The synthetic data generation unit generates synthetic data based on the simulation of geological processes and disaster occurrence mechanisms. The synthetic data is provided to the model training and optimization unit. The model training and optimization unit uses real data, adversarial samples and synthetic data to train and optimize the prediction model. The data processing module interface unit is responsible for interacting with the data processing module to obtain real data and adversarial samples.

[0024] Furthermore, the synthetic data generation unit generates synthetic data based on the simulation of geological processes and disaster occurrence mechanisms in the following specific steps:

[0025] Understand geological processes: Collect geological data through geological surveys, geophysical exploration, geological drilling, etc., and use geographic information systems (GIS) for spatial analysis and visualization;

[0026] Establish a disaster mechanism model: Based on the understanding of geological processes, use the finite element method to establish a mathematical model to simulate the disaster mechanism;

[0027] Generate simulated data: Use the established disaster mechanism model to generate synthetic data that is similar to real data but more numerous;

[0028] Validate and adjust the model: Use statistical methods (such as histograms, scatter plots, correlation analysis, etc.) to compare the distribution and relationship of synthetic data and real data. If there are differences, the parameters or structure of the model can be adjusted to improve accuracy.

[0029] Applied to predictive model training: Synthetic data is input into the predictive model as part of the training set and trained together with real data, which can improve the generalization ability and accuracy of the model.

[0030] Furthermore, the model training and optimization unit trains the prediction model in the following specific steps:

[0031] LSTM modeling of time series geological data:

[0032] Data preprocessing: cleaning time series geological data, removing outliers and missing values, and standardizing or normalizing the data to make them on the same scale;

[0033] LSTM model construction: Define an LSTM network structure, including an input layer, an LSTM layer (possibly containing multiple layers), a fully connected layer, and an output layer. The input layer is responsible for receiving time series geological data, the LSTM layer is responsible for capturing long-term dependencies in the data, the fully connected layer converts the output of the LSTM layer into the required feature representation, and the output layer outputs the corresponding prediction results based on task requirements (such as classification or regression).

[0034] Model training: Use the training dataset to train the LSTM model and optimize the model parameters through the back-propagation algorithm. The loss function usually uses mean square error (MSE) or cross entropy loss.

[0035] Model evaluation: Use the validation dataset to evaluate the model's performance, such as accuracy, recall, and F1 score. Adjust the model structure or parameters based on the evaluation results to improve model performance.

[0036] The output of LSTM is used as a feature input to the SVM classifier:

[0037] Feature extraction: Extracting feature representations of the output layer or fully connected layer from the trained LSTM model. The feature representations contain information about long-term dependencies and geological change trends in time series geological data.

[0038] SVM model construction: Define an SVM classifier, including a kernel function (such as a linear kernel, RBF kernel, etc.) and a penalty parameter C, use the extracted feature representation as input, and train the SVM classifier;

[0039] Model training and evaluation: Use the training dataset to train the SVM model, and use the validation dataset to evaluate the model performance, such as classification accuracy. Adjust the parameters of the SVM model based on the evaluation results to improve model performance.

[0040] Furthermore, the data security module includes a data encryption unit, an access control unit, a security audit and monitoring unit, a privacy compliance management unit and a data backup and recovery unit. The data encryption unit is responsible for data encryption and decryption, the access control unit is responsible for managing user roles, permissions and authentication, the security audit and monitoring unit is responsible for monitoring system logs, detecting abnormal access and potential threats, the privacy compliance management unit ensures that system operations comply with privacy regulations and processes user privacy data, the data backup and recovery unit is responsible for regular data backup and disaster recovery plans, the data encryption unit encrypts the data before data storage, and then passes the encrypted data to the data center for storage, the access control unit controls access to encrypted data and unencrypted data based on the user's role and permissions, when a user requests access to data, the access control unit first verifies the user's permissions, and then allows or denies the access request, the security audit and monitoring unit monitors the system log in real time, including the access records of the access control unit and the data operation records of the data center, the data backup and recovery unit regularly copies data from the data center and stores it in a secure location.

[0041] The present invention has the following beneficial effects:

[0042] 1. In the present invention, the intelligent analysis module constructs a prediction model based on LSTM and SVM. LSTM is used to capture long-term dependencies in time series geological data and identify geological change trends, while SVM is used to classify or regress geological structures. This combined use gives full play to the advantages of the two algorithms and improves the accuracy and reliability of the prediction model. In addition, the finite element method is used to establish a mathematical model that simulates the mechanism of disaster occurrence. By setting different parameter combinations and initial conditions, the mathematical model is run to generate synthetic data that is similar to the real data but in larger quantities. These synthetic data not only help enrich the training samples and improve the generalization ability of the prediction model, but also provide more data support for in-depth analysis and prediction of geological exploration work.

[0043] 2. In the present invention, by introducing adversarial training technology, the system can generate adversarial samples and improve the robustness of the model to noise and bias. This technology helps to improve the predictive ability and stability of the model in complex geological environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 This is a system block diagram of the geological exploration data intelligent cloud management and analysis system proposed in this invention. DETAILED DESCRIPTION

[0045] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0046] See also Figure 1 As shown, the present invention is a geological exploration data intelligent cloud management and analysis system, comprising:

[0047] Data access module: responsible for supporting automatic or manual import of geological survey data from various sources, including field survey data and sensor data;

[0048] Data management and integration module: Receives data from the data collection and upload module, performs version control, ensures data consistency and traceability, integrates data across projects and domains, and forms a unified data view;

[0049] Data processing module: accepts data provided by the data management and integration module, and is responsible for data cleaning, format conversion, and normalization to ensure data quality. It also introduces adversarial training technology to generate adversarial samples, thereby improving the model's robustness to noise and bias.

[0050] Intelligent Analysis Module: This module uses the finite element method to establish a mathematical model simulating the mechanism of disaster occurrence. By setting different parameter combinations and initial conditions, the mathematical model is run to generate synthetic data that is similar to real data but more abundant. This data is then used to train the prediction model. Simultaneously, the module accesses the dataset provided by the data processing module to train the prediction model and build a prediction model. The prediction model uses LSTM to model time series geological data, capture long-term dependencies in the data, and identify geological change trends. The LSTM output is then input as a feature into the SVM classifier to perform classification or regression prediction of geological structures.

[0051] Report generation and sharing module: Receives the report content after analysis from the intelligent analysis module, generates an exploration report, and supports online preview, download, and sharing of the report under permission control;

[0052] Data security module: responsible for data security and privacy protection during data storage, analysis and processing.

[0053] In one embodiment, for the above-mentioned data access module, the data access module includes a data import unit, a data verification unit, a data upload unit and a log recording unit. The data import unit is responsible for receiving data from various sources and converting it into a format recognizable by the system. The data import unit includes a mobile device APP interface, a web page interface and a sensor interface. The mobile device APP interface interacts with the mobile device APP for data and receives field survey data. The web page interface interacts with the web page for data and receives data uploaded by users. The sensor interface interacts with various sensors (such as GPS, seismic wave sensors, magnetic sensors, etc.) for data and receives real-time or periodically collected data. The data verification unit verifies the imported data to ensure the quality of the data. The data verification unit includes a format verification subunit, an integrity verification subunit and a rationality verification subunit. The format verification subunit checks whether the data format meets the system requirements. The integrity verification subunit checks whether the data is complete, such as whether required fields are missing. The rationality verification subunit checks the rationality of the data based on the common sense and rules of geological exploration, such as whether the GPS coordinates are within a reasonable range and whether the geological description meets the common sense of geology. The data upload unit will The verified data is uploaded to the system. The data upload unit includes a database interface and a cache interface. The database interface interacts with the system's database and stores the data in the corresponding table. To improve data access speed, the cache interface stores the data in a cache (such as Redis). The log recording unit records the process and results of data import, verification, and upload. The log recording unit includes an operation log subunit and an error log subunit. The operation log subunit records the time, operator, operation result, and other information of each data import, verification, and upload operation. The error log subunit records error information and processing results encountered during the operation. The data import unit passes the received data to the data verification unit for verification. The data verification unit returns the verification result (pass / fail) to the data import unit. The data verification unit passes the verified data to the data upload unit for upload. The data upload unit returns the upload result (success / failure) to the data verification unit (or log recording unit, for recording). The data upload unit uploads the verified data to the database or cache for storage. The database or cache returns the storage result (success / failure) to the data upload unit (or log recording unit, for recording).

[0054] In one embodiment, for the above-mentioned data processing module, the data processing module includes a data receiving unit, a data cleaning unit, a format conversion unit, a normalization processing unit and an adversarial training unit. The data receiving unit is responsible for receiving data from the data management and integration module and performing preliminary data verification. The data cleaning unit performs cleaning operations on the received data, including removing duplicate data, filling missing values, correcting erroneous data, etc. The format conversion unit converts the data into an appropriate format as needed for subsequent analysis or model training. The normalization processing unit normalizes the data to ensure that the data is compared and analyzed on a unified scale. The adversarial training unit is responsible for generating adversarial samples.

[0055] In one embodiment, for the above-mentioned adversarial training unit, the specific steps of the adversarial training unit to generate adversarial samples are:

[0056] Compute the gradient:

[0057] For a given input sample x and label y, calculate the gradient of the loss function L(x,y) of the LSTM and SVM classifiers on x with respect to x. The gradient represents the rate of change of the loss function in the input space and is the key to guiding how to adjust the input to maximize the loss function.

[0058] Determine the perturbation direction and magnitude:

[0059] According to the calculated gradient, the direction of the disturbance is determined, and the direction of the disturbance is selected as the sign direction of the gradient, that is, The amplitude of the perturbation is controlled by the hyperparameter ε, which determines the amount of noise added to the input sample. The selection of ε requires a balance between the attack effect and the identifiability of the sample. An ε that is too large causes the sample to become unidentifiable, while an ε that is too small cannot effectively mislead the model.

[0060] Generate adversarial examples:

[0061] According to the determined perturbation direction and amplitude, the adversarial sample x' is generated, and x' is calculated by the following formula:

[0062] In one embodiment, for the above-mentioned intelligent analysis module, the intelligent analysis module includes a synthetic data generation unit, a model training and optimization unit, and a data processing module interface unit. The synthetic data generation unit generates synthetic data based on the simulation of geological processes and disaster occurrence mechanisms, and the synthetic data is provided to the model training and optimization unit. The model training and optimization unit uses real data, adversarial samples and synthetic data to train and optimize the prediction model. The data processing module interface unit is responsible for interacting with the data processing module to obtain real data and adversarial samples.

[0063] In one embodiment, for the synthetic data generation unit, the synthetic data generation unit generates synthetic data based on the simulation of geological processes and disaster occurrence mechanisms in the following specific steps:

[0064] Understand geological processes: Collect geological data through geological surveys, geophysical exploration, geological drilling, etc., and use geographic information systems (GIS) for spatial analysis and visualization;

[0065] Establish a disaster mechanism model: Based on the understanding of geological processes, use the finite element method to establish a mathematical model to simulate the disaster mechanism;

[0066] Let's say we want to simulate the occurrence of an earthquake. We use the finite element method (FEM) to simulate the propagation of seismic waves in the Earth's crust. These models need to take into account parameters such as the elasticity, density, and thickness of the crust, as well as the location and intensity of the earthquake source.

[0067] Generate simulated data: Use the established disaster occurrence mechanism model to generate synthetic data that is similar to the real data but in larger quantities. By setting different parameter combinations and initial conditions, run the model to generate simulated data. These data may include seismic wave data, surface displacement data, soil moisture data, etc.

[0068] In earthquake simulations, different earthquake source locations, intensities, and crustal parameters are set, and then the seismic wave propagation model is run to generate data sets for multiple earthquake events. These data sets can include information such as the arrival time, amplitude, and frequency of the seismic waves.

[0069] Validate and adjust the model: Use statistical methods (such as histograms, scatter plots, correlation analysis, etc.) to compare the distribution and relationship of synthetic data and real data. If there are differences, the parameters or structure of the model can be adjusted to improve accuracy.

[0070] In earthquake simulation, the distribution of synthetic seismic wave data and real seismic wave data in terms of amplitude, frequency, etc. If a large difference is found, the parameters of the seismic wave propagation model (such as the elastic modulus and density of the crust) can be adjusted to improve accuracy.

[0071] Applied to predictive model training: Synthetic data is input into the predictive model as part of the training set and trained together with real data, which can improve the generalization ability and accuracy of the model.

[0072] In earthquake prediction, the generated synthetic seismic wave data is input into the earthquake prediction model as part of the training set. By comparing it with real earthquake data, the model can learn more about seismic wave propagation and earthquake generation mechanisms, thereby improving the accuracy of predictions.

[0073] In one embodiment, for the above-mentioned model training and optimization unit, the specific steps of the model training and optimization unit for training the prediction model are as follows:

[0074] LSTM modeling of time series geological data:

[0075] Data preprocessing: cleaning time series geological data, removing outliers and missing values, and standardizing or normalizing the data to make them on the same scale;

[0076] LSTM model construction: Define an LSTM network structure, including an input layer, an LSTM layer (possibly multiple layers), a fully connected layer, and an output layer. The input layer is responsible for receiving time series geological data, the LSTM layer is responsible for capturing long-term dependencies in the data, the fully connected layer converts the output of the LSTM layer into the required feature representation, and the output layer outputs the corresponding prediction results based on task requirements (such as classification or regression).

[0077] Model training: Use the training dataset to train the LSTM model and optimize the model parameters through the back-propagation algorithm. The loss function usually uses mean square error (MSE) or cross entropy loss.

[0078] Model evaluation: Use the validation dataset to evaluate the model's performance, such as accuracy, recall, and F1 score. Adjust the model structure or parameters based on the evaluation results to improve model performance.

[0079] The output of LSTM is used as a feature input to the SVM classifier:

[0080] Feature extraction: Extract the feature representation of the output layer or fully connected layer from the trained LSTM model. The feature representation contains the long-term dependencies and geological change trend information in the time series geological data.

[0081] SVM model construction: Define an SVM classifier, including a kernel function (such as a linear kernel, RBF kernel, etc.) and a penalty parameter C, use the extracted feature representation as input, and train the SVM classifier;

[0082] Model training and evaluation: Use the training data set to train the SVM model, and use the validation data set to evaluate the model's performance, such as classification accuracy. Adjust the SVM model parameters based on the evaluation results to improve model performance.

[0083] Assume we have extracted feature representations for time series geological data using an LSTM model. We can now use these feature representations as input to train an SVM classifier to classify geological structures. For example, we can input features such as seismic activity intensity and groundwater level changes into an SVM classifier to determine whether a region is prone to geological hazards such as earthquakes or landslides.

[0084] In one embodiment, for the above-mentioned data security module, the data security module includes a data encryption unit, an access control unit, a security audit and monitoring unit, a privacy compliance management unit, and a data backup and recovery unit. The data encryption unit is responsible for encrypting and decrypting data, the access control unit is responsible for managing user roles, permissions and authentication, the security audit and monitoring unit is responsible for monitoring system logs, detecting abnormal access and potential threats, the privacy compliance management unit ensures that system operations comply with privacy regulations and processes user privacy data, the data backup and recovery unit is responsible for regular data backup and disaster recovery plans, before data storage, the data encryption unit encrypts the data and then passes the encrypted data to the data center for storage, the access control unit controls access to encrypted data and unencrypted data based on the user's role and permissions, when a user requests access to data, the access control unit first verifies the user's permissions and then allows or denies the access request, the security audit and monitoring unit monitors the system log in real time, including the access records of the access control unit and the data operation records of the data center, the data backup and recovery unit regularly copies data from the data center and stores it in a secure location.

[0085] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A geological exploration data intelligent cloud management and analysis system, characterized by: include: Data access module: responsible for supporting automatic or manual import of geological survey data from various sources, including field survey data and sensor data; Data management and integration module: Receives data from the data collection and upload module, performs version control, integrates data across projects and fields, and forms a unified data view; Data processing module: accepts data provided by the data management and integration module, and is responsible for data cleaning, format conversion, and normalization. It also introduces adversarial training technology and generates adversarial samples to improve the model's robustness to noise and bias. Intelligent Analysis Module: This module uses the finite element method to establish a mathematical model simulating the mechanism of disaster occurrence. By setting different parameter combinations and initial conditions, the mathematical model is run to generate synthetic data that is similar to real data but more abundant. This data is then used to train the prediction model. Simultaneously, the module accesses the dataset provided by the data processing module to train the prediction model and build a prediction model. The prediction model uses LSTM to model time series geological data, capture long-term dependencies in the data, and identify geological change trends. The LSTM output is then input as a feature into the SVM classifier to perform classification or regression prediction of geological structures. Data security module: responsible for data security and privacy protection during data storage, analysis and processing.

2. A geological exploration data intelligent cloud management and analysis system according to claim 1, characterized in that: The data access module includes a data import unit, a data verification unit, a data upload unit and a log recording unit. The data import unit is responsible for receiving data from various sources and converting it into a format recognizable by the system. The data import unit includes a mobile device APP interface, a web page interface and a sensor interface. The data verification unit verifies the imported data. The data verification unit includes a format verification subunit, an integrity verification subunit and a rationality verification subunit. The data upload unit uploads the verified data to the system. The data upload unit includes a database interface and a cache interface. The log recording unit records the process and results of data import, verification and upload. The log recording unit includes an operation log subunit and an error log subunit. The data import unit passes the received data to the data verification unit for verification. The data verification unit returns the verification result to the data import unit. The data verification unit passes the verified data to the data upload unit for upload. The data upload unit returns the upload result to the data verification unit. The data upload unit uploads the verified data to the database or cache for storage. The database or cache returns the storage result to the data upload unit.

3. A geological exploration data intelligent cloud management and analysis system according to claim 1, characterized in that: The data management and integration module includes a data receiving unit, a version control unit, a data integration unit and a data service unit. The data receiving unit is responsible for receiving data transmitted from the data access and performing preliminary processing. The version control unit performs version management on the received data and records the data change history. The data integration unit is responsible for integrating data across projects and fields to form a unified data view. The data service unit provides data query services for other modules or external systems.

4. A geological exploration data intelligent cloud management and analysis system according to claim 1, characterized in that: The data processing module includes a data receiving unit, a data cleaning unit, a format conversion unit, a normalization processing unit and an adversarial training unit. The data receiving unit is responsible for receiving data from the data management and integration module and performing preliminary data verification. The data cleaning unit cleans the received data. The format conversion unit converts the data into an appropriate format. The normalization processing unit normalizes the data. The adversarial training unit is responsible for generating adversarial samples.

5. A geological exploration data intelligent cloud management and analysis system according to claim 4, characterized in that: The specific steps of the adversarial training unit to generate adversarial samples are: Compute the gradient: For a given input sample x and label y, calculate the gradient of the loss function L(x,y) of the LSTM and SVM classifiers on x with respect to x respectively; Determine the perturbation direction and magnitude: According to the calculated gradient, the direction of the disturbance is determined, and the direction of the disturbance is selected as the sign direction of the gradient, that is, Generate adversarial examples: According to the determined perturbation direction and amplitude, the adversarial sample x' is generated, and x' is calculated by the following formula: ∈ is a constant.

6. A geological exploration data intelligent cloud management and analysis system according to claim 1, characterized in that: The intelligent analysis module includes a synthetic data generation unit, a model training and optimization unit, and a data processing module interface unit. The synthetic data generation unit generates synthetic data based on the simulation of geological processes and disaster occurrence mechanisms. The synthetic data is provided to the model training and optimization unit. The model training and optimization unit uses real data, adversarial samples and synthetic data to train and optimize the prediction model. The data processing module interface unit is responsible for interacting with the data processing module to obtain real data and adversarial samples.

7. A geological exploration data intelligent cloud management and analysis system according to claim 6, characterized in that: The synthetic data generation unit generates synthetic data based on the simulation of geological processes and disaster mechanisms, and the specific steps are as follows: Understanding geological processes: Collecting geological data and using geographic information systems for spatial analysis and visualization; Establish a disaster mechanism model: Use the finite element method to establish a mathematical model to simulate the disaster mechanism; Generate simulation data: By setting different parameter combinations and initial conditions, run the mathematical model that simulates the disaster occurrence mechanism to generate simulation data; Validate and tune the model: Use statistical methods to compare the distribution and relationships of synthetic data and real data, and if there are differences, adjust the parameters or structure of the model; Applied to predictive model training: Synthetic data is input into the predictive model as part of the training set and trained together with real data.

8. A geological exploration data intelligent cloud management and analysis system according to claim 6, characterized in that: The specific steps of the model training and optimization unit for training the prediction model are as follows: LSTM modeling of time series geological data: Data preprocessing: cleaning time series geological data, removing outliers and missing values, and standardizing or normalizing the data to make them on the same scale; LSTM model construction: Define an LSTM network structure, including an input layer, an LSTM layer, a fully connected layer, and an output layer. The input layer is responsible for receiving time series geological data, the LSTM layer is responsible for capturing long-term dependencies in the data, the fully connected layer converts the output of the LSTM layer into the required feature representation, and the output layer outputs the corresponding prediction results according to task requirements; Model training: Use the training data set to train the LSTM model and optimize the model parameters through the back propagation algorithm; Model evaluation: Use the validation dataset to evaluate the performance of the model and adjust the model structure or parameters based on the evaluation results; The output of LSTM is used as a feature input to the SVM classifier: Feature extraction: Extracting feature representations of the output layer or fully connected layer from the trained LSTM model. The feature representations contain information about long-term dependencies and geological change trends in time series geological data. SVM model construction: Define an SVM classifier, including the kernel function and penalty parameter C, use the extracted feature representation as input, and train the SVM classifier; Model training and evaluation: Use the training dataset to train the SVM model, use the validation dataset to evaluate the model performance, and adjust the parameters of the SVM model based on the evaluation results.

9. The geological exploration data intelligent cloud management and analysis system according to claim 1, characterized in that: The data security module includes a data encryption unit, an access control unit, a security audit and monitoring unit, a privacy compliance management unit and a data backup and recovery unit. The data encryption unit is responsible for data encryption and decryption. The access control unit is responsible for managing user roles, permissions and authentication. The security audit and monitoring unit is responsible for monitoring system logs, detecting abnormal access and potential threats. The privacy compliance management unit ensures that system operations comply with privacy regulations and processes user privacy data. The data backup and recovery unit is responsible for regular data backup and disaster recovery plans. Before data storage, the data encryption unit encrypts the data and then passes the encrypted data to the data center for storage. The access control unit controls access to encrypted data and unencrypted data based on the user's role and permissions. When a user requests access to data, the access control unit first verifies the user's permissions and then allows or denies the access request. The security audit and monitoring unit monitors the system log in real time, including the access records of the access control unit and the data operation records of the data center. The data backup and recovery unit regularly copies data from the data center and stores it in a secure location.

Images