A mining area supervision and monitoring method, system and electronic equipment

Abstract

The present disclosure relates to the field of deep application of artificial intelligence technology, and in particular to a method, system and electronic equipment for supervision and inspection of mining areas. The method includes: performing multimodal perception on the mining area to be monitored, obtaining multimodal monitoring information of the mining area to be monitored, and predicting the hidden danger development trend of the mining area to be monitored based on the multimodal monitoring information, wherein the multimodal monitoring information includes at least one of scene information and data information; determining the law enforcement task set corresponding to the mining area to be monitored based on the multimodal monitoring information and the hidden danger development trend; determining the location information set of the law enforcement task set in the mining area to be monitored, and determining the law enforcement plan corresponding to the law enforcement task set based on the location information set, so as to conduct on-site law enforcement in the mining area to be monitored according to the law enforcement plan. The present disclosure using the above-mentioned scheme can improve the level of intelligent supervision and inspection.

Description

Technical Field

[0001] The present disclosure relates to the field of deep application of artificial intelligence technology, and in particular to a method, system and electronic equipment for mining area supervision and inspection. Background Art

[0002] Supervision and inspection are important means to ensure safe production in coal mines. However, they are limited in the number of supervisory and inspection personnel, the limited experience of law enforcement personnel, and the lack of intelligent law enforcement methods. Current law enforcement still has problems such as inaccurate supervision and enforcement plans, which mainly rely on investigations, spot checks and surprise inspections by law enforcement departments and regional law enforcement personnel. The comparison of materials and the law enforcement process mainly rely on manpower, resulting in omissions and failure to detect problems, large resource consumption, high labor costs and low efficiency. The supervision and law enforcement system cannot enable on-site law enforcement, and the law enforcement results cannot be fair. Summary of the Invention

[0003] The present disclosure aims to solve one of the technical problems in the related art at least to a certain extent.

[0004] To this end, the first purpose of the present disclosure is to propose a mining area supervision and inspection method to improve the level of intelligent supervision and inspection.

[0005] The second objective of the present disclosure is to provide a mining area supervision and inspection system.

[0006] A third objective of the present disclosure is to provide an electronic device.

[0007] A fourth object of the present disclosure is to provide a computer-readable storage medium.

[0008] A fifth object of the present disclosure is to provide a computer program product.

[0009] To achieve the above objectives, the first embodiment of the present disclosure provides a method for monitoring and supervising a mining area, comprising:

[0010] Performing multimodal sensing on the mining area to be monitored to obtain multimodal monitoring information of the mining area to be monitored, and predicting a hidden danger development trend of the mining area to be monitored based on the multimodal monitoring information, wherein the multimodal monitoring information includes at least one of scene information and data information;

[0011] Determining a set of law enforcement tasks corresponding to the mining area to be monitored based on the multimodal monitoring information and the hidden danger development trend;

[0012] Determine the location information set of the law enforcement task set in the mining area to be monitored, and determine the law enforcement plan corresponding to the law enforcement task set based on the location information set, so as to perform on-site law enforcement on the mining area to be monitored according to the law enforcement plan.

[0013] Optionally, performing multimodal sensing on the mining area to be monitored to obtain multimodal monitoring information of the mining area to be monitored includes:

[0014] Performing scene perception on the mining area to be monitored to obtain hidden danger information of the mining area scene;

[0015] Performing data sensing on the mining area data of the mining area to be monitored to obtain hidden danger information of the mining area data;

[0016] Multimodal fusion is performed on the mining area scene hidden danger information and the mining area data hidden danger information to obtain multimodal monitoring information of the mining area to be monitored.

[0017] Optionally, the mining scene hidden danger information includes personnel active warning information, personnel abnormal card information, and environmental parameter abnormality information. The mining scene hidden danger information obtained by performing scene perception on the mining area to be monitored includes:

[0018] Acquire a mining area monitoring image of the mining area scene of the mining area to be monitored, and determine the personnel active warning information according to the mining area monitoring image;

[0019] Obtaining the movement trajectory of personnel and positioning card data in the mining scene, and determining the abnormal card-carrying information of the personnel based on the mining area monitoring image, the movement trajectory of the personnel and the positioning card data;

[0020] Acquire gas monitoring data of the mining scene, and determine the abnormal environmental parameter information based on the gas monitoring data.

[0021] Optionally, the performing data sensing on the mining area data of the mining area to be monitored to obtain the mining area data hidden danger information includes:

[0022] Extracting features from the mining area data of the mining area to be monitored based on a business rule engine to obtain feature data;

[0023] Extracting a subgraph corresponding to the feature data from a knowledge graph corresponding to a coal mine safety production knowledge base;

[0024] According to the sub-graph, the mining area data hidden danger information corresponding to the mining area data is determined.

[0025] Optionally, predicting the hidden danger development trend of the mining area to be monitored based on the multimodal monitoring information includes:

[0026] The multimodal monitoring information is input into a hidden danger development trend prediction model to obtain the hidden danger development trend of the mining area to be monitored output by the hidden danger development trend prediction model.

[0027] Optionally, determining a set of law enforcement tasks corresponding to the mining area to be monitored based on the multimodal monitoring information and the hidden danger development trend includes:

[0028] Case-based reasoning technology is used to generate a set of law enforcement tasks corresponding to the multimodal monitoring information and the hidden danger development trend.

[0029] Optionally, the method further includes:

[0030] During on-site enforcement of the mining area to be monitored according to the enforcement plan, obtaining a current enforcement location, and generating and issuing enforcement prompt information according to the enforcement task set when the current enforcement location matches any location information in the location information set;

[0031] Obtain the on-site law enforcement results corresponding to the law enforcement prompt information, and generate a series of documents corresponding to the on-site law enforcement results.

[0032] Optionally, the method further includes:

[0033] During on-site law enforcement of the monitored mining area according to the law enforcement plan, in response to receiving law enforcement question information, reply information corresponding to the law enforcement question information is generated and issued based on the coal mine supervision and law enforcement expert knowledge base.

[0034] To achieve the above objectives, a second embodiment of the present disclosure provides a mining area supervision and inspection system, comprising:

[0035] a hidden danger analysis agent, configured to perform multimodal perception of a mining area to be monitored, obtain multimodal monitoring information of the mining area to be monitored, and predict a hidden danger development trend of the mining area to be monitored based on the multimodal monitoring information, wherein the multimodal monitoring information includes at least one of scene information and data information;

[0036] An enforcement plan arrangement agent is used to determine a set of enforcement tasks corresponding to the mining area to be monitored based on the multimodal monitoring information and the hidden danger development trend;

[0037] The law enforcement plan orchestration intelligent agent is also used to determine the location information set of the law enforcement task set in the mining area to be monitored, and determine the law enforcement plan corresponding to the law enforcement task set based on the location information set, so as to perform on-site law enforcement in the mining area to be monitored according to the law enforcement plan.

[0038] To achieve the above-mentioned objectives, a third embodiment of the present disclosure provides an electronic device, including:

[0039] a memory for storing executable program code;

[0040] The processor is used to call and run the executable program code from the memory, so that the electronic device executes the method shown in any one of the first aspects above.

[0041] To achieve the above-mentioned purpose, the fourth aspect of the present disclosure provides a computer-readable storage medium, which stores a computer program. When the computer program is executed, it implements the method shown in any one of the above-mentioned first aspects.

[0042] To achieve the above-mentioned objectives, an embodiment of the fifth aspect of the present disclosure proposes a computer program product, including a computer program, which implements the method shown in any one of the above-mentioned first aspects when executed by a processor.

[0043] In summary, the method, system, and electronic device provided by the present disclosure obtain multimodal monitoring information of the monitored mining area through multimodal sensing of the monitored mining area, and predict the development trend of hidden dangers in the monitored mining area based on the multimodal monitoring information; therefore, safety hazard inspections can be sunk to the mining area data layer, and hidden major safety hazards can be gradually revealed through multimodal multivariate data analysis, solving the problem of the last mile of on-site law enforcement, which is conducive to urging the mining company to actively create safety and prevent major accidents in the mine. Then, based on the multimodal monitoring information and the development trend of hidden dangers, the corresponding law enforcement task set of the monitored mining area is determined; the location information set of the law enforcement task set in the monitored mining area is determined, and the law enforcement plan corresponding to the law enforcement task set is determined based on the location information set, so that on-site law enforcement is carried out in the monitored mining area according to the law enforcement plan; therefore, the problem of precise and differentiated law enforcement can be solved, the frequency of law enforcement visits to the mine can be significantly reduced, and there is a certain degree of help in improving the production efficiency of the mining area, effectively assisting law enforcement personnel to improve law enforcement efficiency, improve the level of intelligence of on-site law enforcement, improve law enforcement efficiency, and improve the effectiveness, usability, and reliability of traditional supervision and inspection systems.

[0044] Additional aspects and advantages of the present disclosure will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The above and / or additional aspects and advantages of the present disclosure will become apparent and readily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0046] Figure 1 A flowchart of a mining area supervision and inspection method provided by an embodiment of the present disclosure;

[0047] Figure 2 A schematic diagram of a process for obtaining hidden danger information of mining areas provided by an embodiment of the present disclosure;

[0048] Figure 3A schematic diagram of the structure of a mining area supervision and inspection system provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0049] The following describes in detail embodiments of the present disclosure, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present disclosure, and should not be construed as limiting the present disclosure.

[0050] It should be noted that the current supervisory and law enforcement process mainly has the following problems:

[0051] Problem 1: There is a lack of intelligent inspection methods for the diverse basic data uploaded by mining areas;

[0052] Problem 2: There is no plan for law enforcement inspections or spot checks, making it impossible to accurately implement differentiated inspections;

[0053] Question 3: The law enforcement results lack fairness, the compilation of a series of documents is time-consuming, and human errors are prone to occur.

[0054] Against the backdrop of the continuous evolution of artificial intelligence (AI) technology and the increasing sophistication of intelligent mining operations, the application of artificial intelligence (AI) is advancing from basic scenarios to core business areas. Its development is characterized by a significant shift from single-awareness to multi-dimensional interaction, and from simple identification to complex decision-making. This technological evolution provides a new path for building precise supervision and inspection based on intelligent algorithms, effectively improving the scientific nature and effectiveness of supervision work, and strengthening the objectivity and fairness of the supervision and law enforcement process through data-driven decision-making mechanisms.

[0055] The present disclosure is described in detail below with reference to specific embodiments.

[0056] In the first embodiment, if Figure 1 As shown, Figure 1 This is a flowchart of a mining area supervision and inspection method provided by an embodiment of the present disclosure. The method can be implemented by a computer program and can be run on a system for performing mining area supervision and inspection. The computer program can be integrated into an application or run as a standalone tool application.

[0057] Among them, the mining area supervision and inspection method can be executed by electronic equipment.

[0058] For example, the mining area supervision and inspection method includes the following steps:

[0059] S101, performing multimodal sensing on the mining area to be monitored to obtain multimodal monitoring information of the mining area to be monitored, and predicting the hidden danger development trend of the mining area to be monitored based on the multimodal monitoring information;

[0060] According to some embodiments, the mining area to be monitored refers to a mining area that requires regulatory inspection.

[0061] In some embodiments, multimodal perception refers to the process of forming a comprehensive understanding of an environment or object through multiple perception methods, including but not limited to scene perception and data perception.

[0062] In some embodiments, multimodal monitoring information refers to information obtained after multimodal perception of the mining area to be monitored, and includes but is not limited to at least one of scene information and data information.

[0063] According to some embodiments, the hidden danger development trend of the monitored mining area refers to the dynamic direction in which the potential risks or problems in the monitored mining area may evolve over time, as predicted based on multimodal monitoring information, including the possibility of their expansion, transformation or deterioration.

[0064] S102, determining a set of law enforcement tasks corresponding to the mining area to be monitored based on the multimodal monitoring information and the hidden danger development trend;

[0065] According to some embodiments, the law enforcement task set includes at least one law enforcement task, which refers to a task that requires supervision and inspection in the mining area to be monitored.

[0066] S103, determining the location information set of the law enforcement task set in the mining area to be monitored, and determining the law enforcement plan corresponding to the law enforcement task set according to the location information set, so as to perform on-site law enforcement in the mining area to be monitored according to the law enforcement plan.

[0067] According to some embodiments, the law enforcement tasks in the law enforcement task set correspond one-to-one to the location information in the location information set.

[0068] In some embodiments, the content of the law enforcement plan includes but is not limited to the law enforcement order and law enforcement path corresponding to the law enforcement task set.

[0069] In summary, the method provided in this embodiment obtains multimodal monitoring information of the monitored mining area through multimodal perception of the monitored mining area, and predicts the development trend of hidden dangers in the monitored mining area based on the multimodal monitoring information. Therefore, safety hazard inspections can be sunk to the mining area data layer, and hidden major safety hazards can be gradually revealed through multimodal multivariate data analysis, solving the problem of the last mile of on-site law enforcement, which is conducive to urging mine owners to actively create safety and prevent major accidents in mines. Then, based on the multimodal monitoring information and the development trend of hidden dangers, the corresponding law enforcement task set of the monitored mining area is determined; the location information set of the law enforcement task set in the monitored mining area is determined, and the law enforcement plan corresponding to the law enforcement task set is determined based on the location information set, so that on-site law enforcement is carried out in the monitored mining area according to the law enforcement plan. Therefore, the problem of precise and differentiated law enforcement can be solved, the frequency of law enforcement visits to mines can be significantly reduced, and the production efficiency of mining areas can be improved to a certain extent. It effectively assists law enforcement personnel to improve law enforcement efficiency, improves the level of intelligence of on-site law enforcement, improves law enforcement efficiency, and improves the effectiveness, usability and reliability of traditional supervision and inspection systems.

[0070] Another embodiment of the present disclosure provides a method for monitoring and supervising a mining area, which can be executed by an electronic device.

[0071] For example, the mining area supervision and inspection method may include the following steps:

[0072] S201, performing scene perception on the mining area to be monitored to obtain hidden danger information of the mining area scene;

[0073] According to some embodiments, mining area scenario hidden danger information refers to hidden danger information related to mining scenarios. This mining area scenario hidden danger information includes, but is not limited to, personnel active warning information, personnel abnormally wearing card information, and abnormal environmental parameter information. Therefore, it can address three core scenario issues: personnel active warning information, abnormal environmental parameters, and personnel abnormally wearing card information, to generate law enforcement clues and evidence.

[0074] According to some embodiments, for the personnel active warning information, a mining area monitoring image of the mining area scene to be monitored may be obtained, and the personnel active warning information may be determined based on the mining area monitoring image.

[0075] In some embodiments, a skeleton model of the human figure can be constructed by acquiring mining surveillance images and locating the human figure within them. The motion trajectories of the skeleton model in consecutive frames of the mining surveillance images are then correlated and analyzed to obtain the spatial movement patterns and motion frequencies of the skeleton model. Dynamic behaviors corresponding to the spatial movement patterns and motion frequencies are then matched based on temporal motion features. If the dynamic behavior is a person-initiated warning dynamic behavior, person-initiated warning information corresponding to the dynamic behavior is generated. Therefore, by extracting motion features through human posture estimation and combining it with temporal motion patterns to capture dynamic patterns, accurate behavior recognition can be achieved, thereby improving the accuracy of obtaining person-initiated warning information.

[0076] An example of a person's active warning dynamic behavior is "waving a hand to warn." In this case, the key focus is identifying spatial movement patterns of the hand and elbow in consecutive frames (such as elbow angle changes and wrist trajectory arc). This allows for a human-centric approach focused on active warnings. In underground environments with complex lighting and task occlusion, human hand-waving recognition is performed based on human posture estimation and temporal feature modeling, achieving highly robust and low-latency dynamic behavior assessment.

[0077] Among them, the deep learning model can be used to locate human figures in mining area monitoring images, and the coordinates of key points of joints such as shoulders, elbows, and wrists can be extracted in real time to build a skeleton model of the portrait.

[0078] According to some embodiments, for the abnormal personnel card information, the personnel movement trajectory and positioning card data in the mining scene can be obtained, and the abnormal personnel card information can be determined based on the mining area monitoring image, personnel movement trajectory and positioning card data.

[0079] In some embodiments, this can be accomplished by acquiring mining area surveillance images and detecting the movement trajectory of personnel within those images; acquiring location card data and comparing the movement trajectory with the location card data; and generating abnormal behavior trajectory information corresponding to the movement trajectory and location card data if the movement trajectory does not match. Therefore, by combining the movement trajectory with the location card data to determine abnormal behavior trajectory information, the accuracy of obtaining abnormal behavior trajectory information can be improved.

[0080] When detecting the movement of people in mining area surveillance images, the system can monitor specific areas. This area can be adjusted based on the actual application scenario. For example, this area can be demarcated using virtual electronic fence technology. This ensures the safety and effectiveness of regional management and control.

[0081] By integrating multi-source sensor data with a logical judgment mechanism, the system correlates and matches personnel movement trajectories with location card data. This allows the identification of typical abnormal events, such as unusual card carrying, prolonged detention, abnormal leadership handovers, and abnormal trajectories of key inspection personnel. For example, if a person enters a mine without a card, or if a person remains inside a mine without leaving, an alarm mechanism can be triggered immediately, enabling timely identification and response to violations.

[0082] According to some embodiments, environmental parameter anomaly information can be obtained by acquiring gas monitoring data; predicting the gas concentration time series characteristics corresponding to the gas monitoring data; matching the gas monitoring data with an anomaly pattern library to obtain anomaly samples corresponding to the gas monitoring data; calculating the similarity between the gas concentration time series characteristics and the anomaly samples based on a weighted dynamic time warping (WDTW) algorithm; and determining the environmental parameter anomaly information corresponding to the gas monitoring data based on the similarity. Therefore, by establishing a closed-loop gas monitoring-analysis-early warning process, the real-time and reliable acquisition of environmental parameter anomaly information can be guaranteed. Secondly, by introducing an anomaly matching mechanism based on the anomaly pattern library and WDTW, the interpretability and traceability of the early warning output can be enhanced.

[0083] In some embodiments, after data collection is performed to obtain gas monitoring data, the gas monitoring data can be preprocessed to obtain preprocessed gas monitoring data; then, the preprocessed gas monitoring data can be input into the prediction model and the abnormal model library respectively to obtain the gas concentration time series characteristics and abnormal samples corresponding to the preprocessed gas monitoring data; then, early warning analysis can be performed based on the gas concentration time series characteristics and abnormal samples. During the early warning analysis, the similarity between the gas concentration time series characteristics and the abnormal samples is calculated based on the weighted dynamic time warping algorithm and combined with the position weight adjustment; finally, anomaly identification is performed based on the similarity, the most likely abnormal category is identified, and the result is output.

[0084] The prediction model includes, but is not limited to, a deep learning model. For example, a hybrid model combining a long short-term memory (LSTM) network and an attention mechanism can be used. This hybrid model can use time series data to predict gas concentration data at multiple future moments to obtain time series characteristics of gas concentration.

[0085] An anomaly pattern library can be constructed by injecting public accident datasets and simulated anomalies. This library includes, but is not limited to, methane and carbon monoxide anomaly patterns. The anomaly pattern library is the foundation for anomaly type identification. It is generated through historical data integration, expert annotation, and simulation generation and expansion. It stores a variety of known typical gas concentration anomaly change patterns (time series characteristics).

[0086] Among them, applying the weighted dynamic time warping (WDTW) algorithm to real-time gas time series data, combined with a position weighting strategy, can improve matching accuracy. Specifically, WDTW has excellent temporal consistency feature modeling capabilities. To further highlight the importance of anomaly occurrence areas, it introduces Gaussian-based time position weights and local gradient change weights. This allows for greater focus on mutation regions and key nodes when calculating the similarity between gas concentration sequences and known anomaly samples. This ensures flexible matching of asynchronous time series while increasing sensitivity to anomaly types. It is suitable for high-frequency gas time series data collected in real-time scenarios and can effectively assist in anomaly tracing and type determination.

[0087] When anomaly identification is performed based on similarity and the most likely anomaly category is determined, the WDTW algorithm can be used to match gas detection data with the anomaly model library data. This algorithm excels at identifying similarities between two time series. This algorithm can measure the weighted DTW distance between the gas concentration series (used for real-time monitoring and prediction) and each anomaly pattern in the anomaly model library. Smaller distances indicate higher similarity. Ultimately, the anomaly pattern with the smallest distance is selected as the type of the current gas concentration change and output in conjunction with the warning level.

[0088] For example, for carbon monoxide monitoring data, deep learning models can be used to mine the temporal characteristics of methane concentrations for prediction based on data processing. Statistical analysis can be used to construct dynamic thresholds for graded early warning. Pattern matching algorithms can then be used to search and compare against a library of carbon monoxide anomaly patterns to identify anomaly types and issue early warnings. Furthermore, a sliding window statistical method can be used to analyze the distribution of historical carbon monoxide concentration data, extracting the 85th and 95th percentiles as dynamic thresholds for first- and second-level warnings. This, combined with fixed alarm thresholds from relevant coal mine safety regulations, establishes a three-level early warning standard, and constructs a three-level early warning model based on dynamic grading. Furthermore, an anomaly pattern matching mechanism is introduced, leveraging the established library of carbon monoxide anomaly patterns to perform similarity comparisons on real-time data. This comparison utilizes an improved dynamic time warping (DTW) method to enhance the ability to recognize nonlinear time series changes, thereby enabling accurate identification and graded early warnings for different types of anomalies (such as drilling operations and roof collapse).

[0089] For example, for methane monitoring data, data smoothing techniques can be used to process raw data. A hybrid model of long short-term memory (LSTM) and attention mechanisms can be used to understand the temporal dependencies of carbon monoxide concentrations for prediction. Dynamic confidence intervals and the duration of rising concentration trends can be used to set graded warnings. An improved pattern matching algorithm can be used to identify anomaly categories within a methane anomaly pattern library and issue warnings. Furthermore, the model inputs a fixed-length historical methane concentration sequence and outputs predicted concentrations for multiple steps into the future. Based on the predicted results, dynamic confidence intervals are established to quantify the uncertainty of the prediction. The duration of the sustained rise in methane concentration is then used as the trigger for graded warnings, enabling the timely identification and grading of abnormal changes.

[0090] Among them, the LSTM unit can effectively capture the long-term and short-term dynamic changes in the concentration series through its gating mechanism, solving the gradient vanishing problem of traditional recurrent neural networks; the attention mechanism gives the model weighted attention to key time steps, especially the information before and after abnormal concentration mutations, which can improve the accuracy and robustness of predictions.

[0091] It should be noted that for abnormal information on environmental parameters, by combining gas monitoring data with time series prediction models, dynamic threshold analysis and abnormal pattern recognition technology, accurate prediction of gas / carbon monoxide concentration trends and multi-level early warning judgments can be achieved.

[0092] S202, performing data sensing on the mining area data to be monitored to obtain hidden danger information of the mining area data;

[0093] According to some embodiments, feature extraction can be performed on the mining area data of the monitored mining area based on a business rule engine to obtain feature data; a subgraph corresponding to the feature data is extracted from the knowledge graph corresponding to the coal mine safety production knowledge base; and based on the subgraph, the mining area data hidden danger information corresponding to the mining area data is determined.

[0094] In some embodiments, the mining area data may be, for example, multivariate heterogeneous data uploaded by the mining area, including but not limited to table data, image data, text data, and other data.

[0095] According to some embodiments, Figure 2 This is a flow chart of obtaining hidden danger information of mining area provided by the embodiment of the present disclosure. Figure 2As shown, this embodiment can be based on the coal mine safety knowledge enhancement framework of Graph RAG, based on the coal mine safety production knowledge base, and combined with major hidden danger identification standards and law enforcement cases to build a business rule engine. By building a collaborative mechanism between the domain knowledge graph and the large language model, semantic retrieval to the causal reasoning process is realized, thereby realizing one-click analysis and early warning of the content risks of the multi-dimensional heterogeneous basic data uploaded by the mining area, and solving the problems of weak knowledge correlation and insufficient logical reasoning ability of traditional RAG technology.

[0096] After acquiring multivariate heterogeneous data, data processing can be performed on the data and stored in the data report data pool. During data processing, the data can be structured to obtain structured multivariate heterogeneous data. This allows for a standardized review of the mining area's basic data to ensure data validity and consistency. Secondly, the structured multivariate heterogeneous data can be aligned, labeled, and classified, and finally stored in the data report data pool.

[0097] When checking the data in the data report data pool, the data to be checked can be put into the mining area data set to be checked, and then the mining area data set to be checked (data set to be checked) is extracted and / or checked.

[0098] Among them, after obtaining structured multi-dimensional heterogeneous data, indicators can be extracted from the structured multi-dimensional heterogeneous data, and data such as table, word, PDF, picture and other types of file data, key features and related features can be extracted; operations such as related feature table operation and maintenance, document attribution analysis, and document legality verification can also be performed.

[0099] For example, a business rule engine can be controlled to extract features from structured, multi-dimensional, heterogeneous data, generating corresponding feature data. A business rule engine refers to business rules / workflows solidified based on the criteria for determining major hidden dangers and the experience of law enforcement personnel. For example, overcapacity production is primarily judged based on annual approved production capacity and monthly output, which can be used to infer whether there are violations. Another example is abnormal withdrawal of safety production expenses.

[0100] Among them, the business rule engine can obtain structured multivariate heterogeneous data from the mining area data set to be inspected, and then perform feature extraction on it.

[0101] In some embodiments, before extracting the subgraph corresponding to the feature data from the knowledge graph corresponding to the coal mine safety production knowledge base, it is also necessary to obtain the regulatory data in the coal mine safety production knowledge base; extract the entities and associated production features in the regulatory data; and map the associated production features to the relationships between different entities in the initial knowledge graph to obtain the knowledge graph corresponding to the coal mine safety production knowledge base. The associated production features refer to regulatory features.

[0102] The Coal Mine Safety Knowledge Base refers to a database related to the coal mine safety industry. A tailor-made "smart toolbox" for supervisory and inspection agencies, it encompasses core data such as national laws and regulations, industry standards, risk and hazard investigation lists, historical accident cases, law enforcement cases, mine area records and reports, diagrams, emergency plans, and expert experience. It transforms scattered safety information into a practical guide for immediate reference.

[0103] For example, the source documents in the regulatory data can be automatically structured and the regulatory features in the source documents can be automatically extracted through the large-scale model prompt word engineering, and then mapped to the entities and the relationships between entities in the knowledge graph, thereby realizing the automatic construction of the knowledge graph.

[0104] In some embodiments, when extracting a subgraph corresponding to feature data from the knowledge graph corresponding to the coal mine safety production knowledge base, the feature data and the knowledge graph can be matched for word space similarity to obtain the subgraph corresponding to the feature data, thereby obtaining the judgment logic and basis for violation identification.

[0105] Among them, the retriever can be configured according to the extracted entities and related production features, word space similarity matching can be performed based on knowledge graph embedding technology, and the related subgraph can be returned. Then, the business rule engine can be combined to judge the existence of hidden dangers and feedback on the basis of violations.

[0106] S203, performing multimodal fusion on the mining area scene hidden danger information and the mining area data hidden danger information to obtain multimodal monitoring information of the mining area to be monitored;

[0107] According to some embodiments, when performing multimodal fusion on mining scene hidden danger information and mining data hidden danger information, the multimodal fusion methods include but are not limited to data-level fusion, feature-level fusion, target-level fusion, etc.

[0108] S204, inputting the multimodal monitoring information into a hidden danger development trend prediction model to obtain a hidden danger development trend of the mining area to be monitored output by the hidden danger development trend prediction model;

[0109] In some embodiments, the hidden danger development trend prediction model can be a model. In this case, when performing multimodal fusion on the mining scene hidden danger information and the mining data hidden danger information, the feature vectors of different modalities can be fused into a unified representation through splicing, weighted averaging or attention mechanism, and then the unified representation is input into the hidden danger development trend prediction model to obtain the hidden danger development trend of the monitored mining area output by the hidden danger development trend prediction model.

[0110] In some embodiments, the hidden danger development trend prediction model may also include multiple single-modal prediction models, which perform separate predictions for each mode in the multi-modal monitoring information to obtain single-modal prediction results for each mode, and finally output the final hidden danger development trend through voting, weighting or Bayesian fusion.

[0111] According to some embodiments, the model types of the hidden danger development trend prediction model include, but are not limited to, spatiotemporal prediction models, multi-task learning models, etc. The spatiotemporal prediction model can be combined with ConvLSTM or spatiotemporal Transformer to handle dynamic changes, thereby improving the accuracy of hidden danger development trend prediction.

[0112] In some embodiments, the initial hidden danger development trend prediction model may be trained through supervised learning using historical hidden danger evolution data as a training set to obtain a hidden danger development trend prediction model.

[0113] S205, using case-based reasoning technology to generate a set of law enforcement tasks corresponding to multimodal monitoring information and hidden danger development trends;

[0114] According to some embodiments, case-based reasoning (CBR) is an experience-based artificial intelligence method, the core idea of ​​which is to solve new problems by retrieving and reusing solutions to similar problems in the past.

[0115] In some embodiments, a case library can be constructed to match multimodal monitoring information and hidden danger development trends in the case library to obtain a set of law enforcement tasks. The case library can be constructed by collecting historical multimodal monitoring information, historical hidden danger development trends, and historical law enforcement tasks.

[0116] In some embodiments, multimodal monitoring information and a set of law enforcement tasks corresponding to hidden danger development trends can be obtained from a case library through similarity matching.

[0117] S206, determining a location information set of the law enforcement task set in the mining area to be monitored, and determining an enforcement plan corresponding to the law enforcement task set based on the location information set;

[0118] According to some embodiments, the enforcement path corresponding to the location information set may be determined with the goal of minimizing the enforcement path, so as to obtain an enforcement plan corresponding to the enforcement path.

[0119] In some embodiments, the task priority corresponding to each law enforcement task in the law enforcement task set can also be determined, and the law enforcement path can be planned by comprehensively considering the task priority and location information to obtain the law enforcement plan corresponding to the law enforcement task set.

[0120] According to some embodiments, corresponding law enforcement plans can be automatically generated by applying the semantic understanding and planning capabilities of the large model and combining them with a set of law enforcement tasks.

[0121] In some embodiments, it is also possible to support law enforcement personnel to add, delete, modify and check the generated law enforcement plans online. Therefore, it can solve the problem of precise and differentiated law enforcement and reduce the frequency of inspections.

[0122] S207, conduct on-site law enforcement in the monitored mining area according to the law enforcement plan;

[0123] According to some embodiments, during on-site enforcement of a monitored mining area according to an enforcement plan, the current enforcement location can be obtained. If the current enforcement location matches any location information in a location information set, an enforcement prompt message is generated and issued based on the enforcement task set. This can improve enforcement quality and address missed inspections.

[0124] In some embodiments, law enforcement prompt information includes but is not limited to content, methods, and steps that require key monitoring.

[0125] According to some embodiments, during the on-site enforcement of the monitored mining area according to the enforcement plan, in response to receiving enforcement question information, reply information corresponding to the enforcement question information can be generated and issued based on the coal mine supervision and enforcement expert knowledge base.

[0126] According to some embodiments, large model distillation technology can be applied to deploy the large model to the underground terminal mobile phone through a lightweight deep neural network engine (Mobile Neural Network, MNN). It can be used offline for weak network / no network scenarios underground. After connecting to the network, the data is automatically synchronized, solving the problem of operation interruption caused by network during on-site law enforcement in the monitored mining area.

[0127] Among them, when law enforcement personnel use the underground terminal mobile phone to conduct on-site law enforcement in the monitored mining area, they can send law enforcement prompt information corresponding to the current law enforcement location through the headset connected to the underground terminal mobile phone.

[0128] Among them, for various types of knowledge about hidden dangers standards and regulations that law enforcement personnel have doubts about or need to verify, the underground terminal mobile phone can provide offline spoken intelligent question-and-answer services, and provide citation sources, providing on-site knowledge support to law enforcement personnel at any time.

[0129] Among them, the underground terminal mobile phone can also support various evidence collection methods such as taking photos, videos, and text, linking the content and progress of specific law enforcement plans to form a traceable electronic evidence chain, meeting the law enforcement evidence collection needs in complex scenarios, and thus generating corresponding on-site law enforcement results.

[0130] S208, obtaining the on-site law enforcement results corresponding to the law enforcement prompt information, and generating a series of documents corresponding to the on-site law enforcement results.

[0131] According to some embodiments, based on on-site law enforcement results and historical fair law enforcement cases, the large-scale model logical reasoning and aggregation capabilities can be combined with structured case information and document templates to automatically complete the generation of a series of document contents such as on-site records, law enforcement documents, case handling reports, and case filing decisions.

[0132] In summary, the method provided in this embodiment, therefore, constructs a mining supervision and inspection process by following the design concept of "data-driven, knowledge-enhanced, intelligent decision-making, and closed-loop response". The overall technical route is based on the expert knowledge base in the field of coal mine safety production, and applies dynamic knowledge graphs, large language models and multi-agent collaborative technology. It integrates multimodal perception, knowledge fusion, intelligent analysis and automatic response, which can enable the intelligence of on-site supervision and inspection law enforcement, and provide full-process empowerment for the intelligence of the law enforcement process of mine safety supervision and inspection law enforcement personnel.

[0133] In order to implement the above embodiments, the present disclosure also proposes a mining area supervision and inspection system.

[0134] For example, Figure 3 This is a schematic diagram of the structure of a mining area supervision and inspection system provided by the embodiment of the present disclosure. Figure 3 As shown, the mining supervision and inspection system includes an agent center, which includes:

[0135] The hidden danger analysis agent is used to perform multimodal perception of the monitored mining area, obtain multimodal monitoring information of the monitored mining area, and predict the hidden danger development trend of the monitored mining area based on the multimodal monitoring information, wherein the multimodal monitoring information includes at least one of scene information and data information;

[0136] The enforcement plan arrangement agent is used to determine the corresponding enforcement task set for the mining area to be monitored based on multimodal monitoring information and hidden danger development trends;

[0137] The enforcement plan arrangement agent is also used to determine the location information set of the enforcement task set in the mining area to be monitored, and determine the enforcement plan corresponding to the enforcement task set based on the location information set, so as to perform on-site enforcement in the mining area to be monitored according to the enforcement plan.

[0138] Optionally, the mining area supervision and inspection system also includes:

[0139] The task decomposer is used to decompose the supervision and monitoring information into tasks to determine the supervision and monitoring tasks corresponding to the supervision and monitoring information.

[0140] The supervisory monitoring information can be, for example, tasks and / or questions entered by law enforcement personnel. It can be text or spoken. All functions can be accessed through a single portal and displayed uniformly via voice or natural language.

[0141] The task decomposer can decompose the monitoring information based on semantic understanding and divide it into tasks. This task decomposer can be performed by a large model (such as DeepSeek).

[0142] Among them, the supervision and monitoring task is any one of the safety hazard analysis task, the law enforcement plan preparation task Task3, the on-site law enforcement task Task4, and the law enforcement document generation task Task5.

[0143] The Safety Hazard Analysis task is used to initiate a safety hazard analysis of a mining area. This task includes Task 1, a one-click safety hazard analysis task, and Task 2, a basic mining area data analysis and early warning task. Before enforcing a law, law enforcement officers can initiate the Safety Hazard Analysis task through voice and / or text messages, automatically triggering Task 1 and Task 2.

[0144] Among them, the enforcement plan preparation task Task3 is used to instruct the preparation of the enforcement plan.

[0145] Among them, on-site law enforcement task Task 4 is used to indicate that law enforcement personnel are conducting on-site law enforcement in the mining area and need law enforcement assistance.

[0146] Among them, the law enforcement document generation task Task5 is used to instruct the generation of a series of documents.

[0147] Optionally, the hidden danger analysis agent includes an intelligent assessment agent and an intelligent text analysis agent.

[0148] Among them, supervision and monitoring tasks correspond one-to-one with the intelligent agents in the Agent Center. For example, the intelligent analysis agent is used to perform Task 1, the intelligent text analysis agent is used to perform Task 2, the law enforcement plan arrangement agent is used to perform Task 3, the law enforcement assistant agent is used to perform Task 4, and the text generation agent is used to perform Task 5.

[0149] The Intelligent Assessment Agent can remotely assess mining risks by combining visual and data analysis. Specifically, it can perform a one-click inspection of various monitoring systems and scenario-specific algorithms for targeted law enforcement units, ultimately identifying and summarizing all potential safety hazards.

[0150] Optionally, the mining supervision and inspection system also includes a global parameter transfer agent. The output results of each intelligent agent in the Agent Center, that is, key information, can be transferred to other tasks through the global parameter transfer agent to achieve the overall goal.

[0151] Optionally, the mining area supervision and inspection system also includes:

[0152] Enhanced Retrieval Agent (RAG Agent) is used to search the knowledge base, solve knowledge limitations and illusion problems, improve the ability to handle complex tasks, and optimize output quality and professionalism.

[0153] Optionally, the mining area supervision and inspection system also includes:

[0154] The prompt word transmitter Agent is used to transmit detailed descriptions of tasks, related information, and the output results of the previous task to each agent in the form of prompts, ensuring that the agent receives accurate and complete input information, thereby guiding the agent to perform tasks according to the expected goals and improving the accuracy and efficiency of task processing.

[0155] Optionally, the mining supervision and inspection system can also summarize the output of each agent in the Agent Center for law enforcement personnel to provide evaluation and feedback, and assist in the iterative upgrade of the agents.

[0156] Among them, the mining area supervision and inspection system adopts a "planning-execution-tool-verification" architecture, with the task decomposer as planning, each task as execution, the Agent Center as a packaged tool, and the summary output as verification. Therefore, it can use artificial intelligence technology to improve the intelligence of supervision and law enforcement, solve the problem of the last mile of on-site law enforcement, and effectively assist law enforcement personnel to improve law enforcement efficiency.

[0157] Optionally, before enforcing the law, law enforcement officers can initiate a one-click safety hazard analysis task, which automatically calls the intelligent research and judgment and intelligent text analysis agents to obtain hazard information for the corresponding mining area. Afterwards, they can initiate an enforcement plan compilation task, automatically calling the plan arrangement agent to formulate a corresponding enforcement plan based on the aforementioned inspection content, while also supporting online editing. Then, when law enforcement officers arrive at the scene, they can initiate an on-site enforcement task, which automatically calls the enforcement assistant agent. It will automatically remind them of what to check, the inspection steps, the inspection basis, and the on-site evidence based on the corresponding location; at the same time, they can understand real-time problems and solutions through intelligent Q&A. After the enforcement is completed, they can initiate an enforcement document generation task, which automatically calls the document generation agent. Based on the on-site enforcement and evidence, it automatically generates various documents such as on-site records, enforcement documents, and penalty decisions, and evaluates the enforcement agent's assistance process.

[0158] It should be noted that the aforementioned explanation of the embodiment of the mining area supervision and inspection method is also applicable to the mining area supervision and inspection system of this embodiment, and will not be repeated here.

[0159] In summary, the system provided by the embodiment of the present disclosure can sink the safety hazard inspection to the mining area data layer through the application of intelligent bodies, and the hidden major safety hazards can be gradually revealed through multimodal and multivariate data analysis, solving the problem of the last mile of on-site law enforcement, which is conducive to urging the mining parties to take the initiative to create safety and prevent major accidents in mines; secondly, the application of intelligent bodies can solve the problem of precise and differentiated law enforcement, significantly reduce the frequency of law enforcement visits to mines, and provide certain help to improve the production efficiency of mining areas, and effectively assist law enforcement personnel to improve law enforcement efficiency; in addition, by applying intelligent bodies to empower the supervision and law enforcement business process, the intelligence of supervision and inspection can be improved based on artificial intelligence technology, the level of intelligence of on-site law enforcement can be improved, the efficiency of law enforcement can be improved, and the effectiveness, usability and reliability of traditional supervision and inspection systems can be improved.

[0160] In order to implement the above embodiments, the present disclosure also proposes an electronic device, including: a processor, and a memory communicatively connected to the processor; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory to implement the method provided by the above embodiments.

[0161] The electronic device includes but is not limited to edge devices, body cameras, and other devices, and the edge devices include but are not limited to mobile phones, computers, and other devices. Deploying the method provided in the above embodiment to the electronic device can achieve lightweight edge deployment of large models.

[0162] In order to implement the above embodiments, the present disclosure further proposes a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the methods provided by the above embodiments.

[0163] In order to implement the above embodiments, the present disclosure further provides a computer program product, including a computer program, which implements the methods provided in the above embodiments when executed by a processor.

[0164] The collection, storage, use, processing, transmission, provision and disclosure of user personal information involved in this disclosure are in compliance with relevant laws and regulations and do not violate public order and good morals.

[0165] It is important to note that personal information collected from users should be used for legitimate and reasonable purposes and should not be shared or sold beyond these legitimate uses. Furthermore, such collection / sharing should be conducted only after receiving the user's informed consent, including but not limited to notifying the user to read the user agreement / user notice and sign an agreement / authorization that includes the relevant user information before using the feature. Furthermore, any necessary steps must be taken to safeguard and secure access to such personal information and ensure that others with access to personal information comply with its privacy policy and procedures.

[0166] This disclosure contemplates providing implementations that allow users to selectively block the use or access of personal information data. Specifically, this disclosure contemplates providing hardware and / or software to prevent or block access to such personal information data. Risks can be minimized by limiting data collection and deleting data once it is no longer needed. Furthermore, where applicable, such personal information can be de-identified to protect user privacy.

[0167] In the descriptions of the aforementioned embodiments, the reference terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" mean that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present disclosure. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and features of different embodiments or examples without contradiction.

[0168] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. Throughout the present disclosure, "plurality" means at least two, such as two, three, etc., unless otherwise specifically defined.

[0169] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code comprising one or more executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present disclosure includes additional implementations in which functions may be performed out of the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present disclosure belong.

[0170] The logic and / or steps represented in a flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM) or flash memory, fiber optic devices, and portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting, or processing it in another suitable manner as necessary, and then storing it in a computer memory.

[0171] It should be understood that various parts of the present disclosure can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having logic gate circuits for implementing logical functions on data signals, an application-specific integrated circuit having suitable combinational logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0172] Those skilled in the art will appreciate that all or part of the steps in the method for implementing the above-mentioned embodiment can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.

[0173] In addition, the functional units in the various embodiments of the present disclosure may be integrated into a single processing module, each unit may exist physically separately, or two or more units may be integrated into a single module. The aforementioned integrated modules may be implemented in the form of hardware or in the form of software functional modules. If the integrated modules are implemented in the form of software functional modules and sold or used as independent products, they may also be stored in a computer-readable storage medium.

[0174] The storage medium mentioned above may be a read-only memory, a magnetic disk, or an optical disk, etc. Although the embodiments of the present disclosure have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present disclosure. A person of ordinary skill in the art may make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present disclosure.

Claims

1. A mining area supervision and inspection method, characterized in that: include: Performing multimodal sensing on the mining area to be monitored to obtain multimodal monitoring information of the mining area to be monitored, and predicting a hidden danger development trend of the mining area to be monitored based on the multimodal monitoring information, wherein the multimodal monitoring information includes at least one of scene information and data information; Determining a set of law enforcement tasks corresponding to the mining area to be monitored based on the multimodal monitoring information and the hidden danger development trend; Determining a location information set of the law enforcement task set in the mining area to be monitored, and determining an enforcement plan corresponding to the law enforcement task set based on the location information set, so as to perform on-site law enforcement in the mining area to be monitored according to the enforcement plan; The multimodal sensing of the mining area to be monitored to obtain multimodal monitoring information of the mining area to be monitored includes: Performing scene perception on the mining area to be monitored to obtain hidden danger information of the mining area scene; Performing data sensing on the mining area data of the mining area to be monitored to obtain hidden danger information of the mining area data; Performing multimodal fusion on the mining area scene hidden danger information and the mining area data hidden danger information to obtain multimodal monitoring information of the mining area to be monitored; The data sensing of the mining area data of the mining area to be monitored to obtain the hidden danger information of the mining area data includes: Extracting features from the mining area data of the monitored mining area based on a business rule engine to obtain feature data, wherein the business rule engine is constructed based on a coal mine safety production knowledge base, combined with major hidden danger identification standards and law enforcement cases, and solidified business rules / workflows based on major hidden danger determination standards and the experience of law enforcement personnel; Extracting a subgraph corresponding to the feature data from a knowledge graph corresponding to a coal mine safety production knowledge base; According to the sub-graph, the mining area data hidden danger information corresponding to the mining area data is determined.

2. The method according to claim 1, characterized in that The mining scene hidden danger information includes personnel active warning information, personnel abnormal card information and environmental parameter abnormal information. The mining scene hidden danger information obtained by performing scene perception on the mining area to be monitored includes: Acquire a mining area monitoring image of the mining area scene of the mining area to be monitored, and determine the personnel active warning information according to the mining area monitoring image; Obtaining the movement trajectory of personnel and positioning card data in the mining scene, and determining the abnormal card-carrying information of the personnel based on the mining area monitoring image, the movement trajectory of the personnel and the positioning card data; Acquire gas monitoring data of the mining scene, and determine the abnormal environmental parameter information based on the gas monitoring data.

3. The method according to claim 1, characterized in that The predicting of the hidden danger development trend of the mining area to be monitored based on the multimodal monitoring information includes: The multimodal monitoring information is input into the hidden danger development trend prediction model to obtain the hidden danger development trend of the monitored mining area output by the hidden danger development trend prediction model, wherein the hidden danger development trend prediction model is one model, or includes multiple single-modal prediction models; the model types of the hidden danger development trend prediction model include spatiotemporal prediction models and multi-task learning models; the spatiotemporal prediction model is combined with ConvLSTM or spatiotemporal Transformer to process dynamic changes, and the hidden danger development trend prediction model is obtained by supervised learning training of the initial hidden danger development trend prediction model using historical hidden danger evolution data as a training set.

4. The method according to claim 1, wherein Determining a set of law enforcement tasks corresponding to the mining area to be monitored based on the multimodal monitoring information and the hidden danger development trend includes: Case-based reasoning technology is used to generate a set of law enforcement tasks corresponding to the multimodal monitoring information and the hidden danger development trend.

5. The method according to claim 1, wherein The method further comprises: During on-site enforcement of the mining area to be monitored according to the enforcement plan, obtaining a current enforcement location, and generating and issuing enforcement prompt information according to the enforcement task set when the current enforcement location matches any location information in the location information set; Obtain the on-site law enforcement results corresponding to the law enforcement prompt information, and generate a series of documents corresponding to the on-site law enforcement results.

6. The method according to claim 1, characterized in that The method further comprises: During on-site law enforcement of the monitored mining area according to the law enforcement plan, in response to receiving law enforcement question information, reply information corresponding to the law enforcement question information is generated and issued based on the coal mine supervision and law enforcement expert knowledge base.

7. A mining area supervision and monitoring system, characterized in that: include: a hidden danger analysis agent, configured to perform multimodal perception of a mining area to be monitored, obtain multimodal monitoring information of the mining area to be monitored, and predict a hidden danger development trend of the mining area to be monitored based on the multimodal monitoring information, wherein the multimodal monitoring information includes at least one of scene information and data information; An enforcement plan arrangement agent is used to determine a set of enforcement tasks corresponding to the mining area to be monitored based on the multimodal monitoring information and the hidden danger development trend; The enforcement plan arrangement agent is further configured to determine a location information set of the enforcement task set in the mining area to be monitored, and determine an enforcement plan corresponding to the enforcement task set based on the location information set, so as to perform on-site enforcement in the mining area to be monitored according to the enforcement plan; The hidden danger analysis intelligent agent is further used to perceive the mining scene of the mining area to be monitored and obtain hidden danger information of the mining scene; Performing data sensing on the mining area data of the mining area to be monitored to obtain hidden danger information of the mining area data; Performing multimodal fusion on the mining area scene hidden danger information and the mining area data hidden danger information to obtain multimodal monitoring information of the mining area to be monitored; The data sensing of the mining area data of the mining area to be monitored to obtain the hidden danger information of the mining area data includes: Extracting features from the mining area data of the monitored mining area based on a business rule engine to obtain feature data, wherein the business rule engine is constructed based on a coal mine safety production knowledge base, combined with major hidden danger identification standards and law enforcement cases, and solidified business rules / workflows based on major hidden danger determination standards and the experience of law enforcement personnel; Extracting a subgraph corresponding to the feature data from a knowledge graph corresponding to a coal mine safety production knowledge base; According to the sub-graph, the mining area data hidden danger information corresponding to the mining area data is determined.

8. An electronic device, characterized in that: The electronic device comprises: a memory for storing executable program code; A processor is configured to call and run the executable program code from the memory, so that the electronic device executes the method according to any one of claims 1 to 6.

Images