Coal mining data management system and method based on video artificial intelligence

Abstract

The invention provides a coal mining data management system and method based on video artificial intelligence, and relates to the field of big data, and the system comprises a data collection module which is used for accessing video data of each scene of a coal mine; the identification module is used for identifying equipment and personnel; the analysis module is used for realizing artificial intelligent scene analysis; the information selection module is used for obtaining selected feature scene data; the comprehensive management module is used for realizing resource sharing and information intercommunication; the control linkage protection module is used for realizing intelligent linkage; and the user interface module provides an operation interface. According to the invention, key elements and dynamic changes thereof in a coal mining scene can be accurately identified and understood, and the fusion condition of virtual reality and a real scene can be more intuitively displayed and analyzed.

Description

Technical Field

[0001] The present invention relates to the field of big data technology, and in particular to a coal mining data management system and method based on video artificial intelligence. Background Art

[0002] With the continuous progress and development of the coal mining industry, the demand for safety management, operational efficiency and intelligent management methods has been raised to a new level. However, for a long time, the coal mining data management system has relied on traditional manual monitoring methods and relatively basic automation technology. This method is not only labor-intensive, but also has monitoring blind spots, making it difficult to achieve comprehensive and all-round monitoring of various scenes in the coal mine, including the complex environment underground and the equipment above the mine. The data collection and update frequency of traditional monitoring systems is limited, and it is impossible to provide real-time scene dynamic information, which often leads to the lack of the latest and most accurate data support for management when making decisions.

[0003] At the same time, manual monitoring is limited by human physiological and psychological factors, such as fatigue and distraction, and is prone to missed judgments and misjudgments, which is extremely disadvantageous in a high-risk industry such as coal mining. Although traditional automation technology has reduced the burden of manual work to a certain extent, due to its limited degree of intelligence, it is unable to autonomously identify and analyze complex situations in the coal mining process, let alone warn of potential safety hazards in advance. This situation has largely restricted the safety and overall efficiency of coal mining, making the coal mining industry face huge safety risks while pursuing high production and high efficiency. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a coal mining data management system and method based on video artificial intelligence, which can accurately identify and understand the key elements and their dynamic changes in the coal mining scene, more intuitively display and analyze the integration of virtual reality and real scenes, more deeply explore and apply the value of feature data, more smoothly realize resource sharing and information exchange, more quickly respond to and deal with emergencies, and more conveniently perform user operations and management.

[0005] In order to solve the above technical problems, the technical solution of the present invention is as follows: In a first aspect, a coal mining data management system based on video artificial intelligence, the system comprising: The data acquisition module is used to access the video data of various scenes in the coal mine, including the data acquisition of equipment above the mine, and store it in the database; The recognition module extracts features and calculates trajectories of devices and personnel in the form of task slots based on videos of different scenes to obtain recognition results; The analysis module is used to combine the recognition results with AR virtual technology enhanced visualization technology, integrate the recognition results, equipment condition data and video streams for analysis, and restore virtual reality through artificial intelligence recognition and AR virtual technology enhanced visualization to achieve artificial intelligence intelligent scene analysis; The information selection module selects feature data based on the results of AI intelligent scene recognition and analysis, and implements multi-level statistics, analysis, and prediction on information of different dimensions to obtain the selected feature scene data; The comprehensive management module is used to integrate video surveillance, feature scene data, and character behavior analysis, and to achieve resource sharing and information exchange through unified coordination of the upper-level comprehensive security management platform; The control linkage protection module is used to perform various interlocking protection controls at the equipment level and information level based on resource sharing and information exchange, and realize intelligent information linkage, including early warning, recording, remote warning and linkage control; The user interface module is used to provide an operation interface, including video viewing, warning information display, system settings, production data visualization, and supports multi-platform access.

[0006] Furthermore, the video data of various scenes in the coal mine is accessed, including the data collection of equipment on the well, and stored in the database, including: Deploy intelligent monitoring equipment in various key scenarios of coal mines, including fully mechanized mining, fully mechanized excavation working faces, key transportation nodes, pump rooms and substation locations; Connect the deployed intelligent monitoring equipment to the network system of the coal mine, develop a data collection interface, and connect the interface with the data management system of the coal mine to obtain a data collection interface; Through the data acquisition interface, the data sent by the intelligent monitoring equipment is collected in real time, including video images, personnel location, personnel behavior, equipment operating conditions, and safety environment factors; Store the data in the database to achieve data storage.

[0007] Furthermore, according to different scene videos, feature extraction and trajectory calculation are performed on equipment and personnel in the form of task slots to obtain recognition results, including: According to the classification of video data, different task slots are created using artificial intelligence technology, and the classified video data are assigned to the corresponding task slots; In each task slot, the features of the equipment and personnel are extracted from the video data and the motion trajectory is calculated by using the feature extraction algorithm and the trajectory calculation algorithm to obtain the results of feature extraction and trajectory calculation; Based on the results of feature extraction and trajectory calculation, the identification results of equipment and personnel are generated, including the operating status of the equipment, fault warning, and the location and behavior information of the personnel.

[0008] Furthermore, in each task slot, the features of the equipment and personnel are extracted from the video data and the motion trajectory is calculated by using the feature extraction algorithm and the trajectory calculation algorithm to obtain the results of the feature extraction and trajectory calculation, including: In each task slot, the classified video data is subjected to feature extraction through a convolutional neural network to obtain key features. The key feature data includes the appearance features, operating status, and behavioral features of the equipment. Obtain the continuous time point data of the feature from the key feature, and pass Calculate the movement trajectory of equipment and personnel, where Pt It's at the time t The predicted location of is the known position at the previous time point t−1, It's at the time t−1 speed, It's at the time t−1 The acceleration of From the time point t−1 To the time point t The time interval K is the correction factor for adjusting the trajectory by external forces and noise, It's at the time t′ The external forces acting on the object, Indicates at a point in time t′ Noise or measurement error, From the time point t−1 arrive t The cumulative effect of external forces and noise on the position of an object.

[0009] Furthermore, the recognition results are combined with AR virtual technology to enhance visualization technology, and the recognition results, equipment condition data and video streams are integrated and analyzed. The virtual reality is restored through artificial intelligence recognition and AR virtual technology to enhance visualization and realize intelligent scene analysis, including: Use artificial intelligence technology to perform scene analysis and behavior prediction on the recognition results to obtain scene understanding results; Based on the recognition and scene understanding results, create virtual models, including 3D representations of devices, objects, and environments; Map the equipment condition data and recognition results to the virtual model, and perform fusion analysis on the AI ​​recognition results, equipment condition data, and video streams to obtain fused data; The fused data is enhanced and displayed in the real scene through AR virtual technology, restoring a virtual scene that is highly integrated with reality; According to real-time data changes, the content and status in the AR virtual technology scene are dynamically updated to achieve intelligent scene analysis.

[0010] Furthermore, based on the results of the artificial intelligence intelligent scene recognition analysis, feature data is selected, and multi-level statistics, analysis and prediction are performed on information of different dimensions to obtain selected feature scene data, including: Evaluate the results of AI intelligent scene recognition analysis to perform dimensionality reduction to select feature data; Based on the selected characteristic data, multi-level statistical analysis is performed, including data distribution, correlation analysis, and cluster analysis; Based on the results of statistical analysis, a prediction model is constructed to predict the scenario data or behavior to obtain the selected feature scenario data.

[0011] Furthermore, based on the selected characteristic data, multi-level statistical analysis is performed, including data distribution, correlation analysis, and cluster analysis, including: According to the selected characteristic data, the mean and variance of the data are calculated to obtain the distribution state of the data; According to the mean and variance of the data, correlation analysis was performed using the Pearson correlation coefficient; According to the selected feature data, Determine the final number of clusters, apply the selected clustering algorithm to perform cluster analysis on the data to obtain the cluster analysis results, where: is the average distance between a data point and other data points in the same cluster, is the minimum average distance between a data point and points in other clusters, Yes and The larger value of It is a cluster The number of data points in is the average size of all clusters, It is a cluster The area or volume of It is a cluster i The number of data points in It is a cluster i The number of data points marked as outliers in It is a cluster i The proportion of outliers in Based on the data distribution status, data correlation analysis and data cluster analysis results, multi-level statistical analysis results can be obtained.

[0012] Furthermore, video surveillance, feature scene data, and character behavior analysis are integrated and coordinated through the upper-level integrated security management platform to achieve resource sharing and information exchange, including: Analyze and plan data and establish a comprehensive security management platform; Connect video surveillance to the integrated security management platform, associate feature scene data with video surveillance, and connect character behavior analysis with the integrated security management platform to obtain an upper-level integrated security management platform; Through the unified coordination of the upper-level comprehensive security management platform, resource sharing and information exchange can be achieved.

[0013] Furthermore, based on resource sharing and information exchange, various interlocking protection controls at the equipment level and information level are carried out to realize intelligent information linkage, including early warning, recording, remote warning and linkage control, including: Design interlocking protection control scheme based on resource sharing and information exchange, and perform interlocking protection control on existing equipment at the equipment level and information level to achieve interlocking protection; Comprehensive testing and verification of the interlocking protection control after implementation is carried out to achieve intelligent linkage.

[0014] In a second aspect, a method for coal mining data management based on video artificial intelligence is provided, the method comprising: Collect video data of various scenes in coal mines, organize and store the collected video data and equipment data in the database; The features of the devices and people in the video are extracted in the form of task slots, and the extracted features and trajectory data are converted into artificial intelligence scene recognition results to obtain artificial intelligence scene recognition results; Based on AR virtual technology enhanced visualization technology, the artificial intelligence scene recognition results, equipment condition data and original video stream are integrated and analyzed, and the virtual reality scene is restored through the combination of artificial intelligence recognition and AR virtual technology to realize artificial intelligence intelligent scene analysis; According to the analysis results of artificial intelligence intelligent scene recognition, key feature data are screened out, and multi-level statistics, analysis and prediction are performed on the key feature data to obtain the selected feature scene data; Integrate video surveillance, feature scene data, and character behavior analysis, and coordinate through the upper-level integrated security management platform to achieve resource sharing and information exchange between subsystems; According to the results of resource sharing and information exchange, interlocking protection control at the equipment level and system level is performed to realize intelligent linkage between various application subsystems, including automatic generation of early warning information, record preservation, remote warning sending and linkage control; According to the intelligent linkage between subsystems, an operation interface is set up to realize the management of monitoring and coal mining.

[0015] The above solution of the present invention includes at least the following beneficial effects: The system can collect video data of various scenes in coal mines in real time, and organize and store these data in the database through efficient data processing processes. By extracting features and calculating trajectories of equipment and personnel in the video in the form of task slots, the system can accurately identify and understand the key elements and their dynamic changes in coal mining scenes. Combined with AR virtual technology enhanced visualization technology, the system can integrate and analyze the recognition results, equipment condition data and original video streams to restore a virtual scene that is highly integrated with reality. The system can filter out key feature data based on the analysis results of artificial intelligence scene recognition, and perform multi-level statistics, analysis and prediction on these data. Through the unified coordination of the upper-level comprehensive safety management platform, the system can integrate video surveillance, feature scene data, and character behavior analysis to achieve resource sharing and information exchange. Based on the results of resource sharing and information exchange, the system can perform interlocking protection control at the equipment level and system level to achieve intelligent linkage between various application subsystems. This includes the automatic generation of early warning information, record keeping, remote warning sending, and linkage control functions, which help to respond to emergencies in a timely manner and ensure the safe and stable operation of coal mining. The system has an intuitive and easy-to-use operation interface, which supports video viewing, warning information display, system settings, production data visualization, and multi-platform access. This helps to reduce the difficulty of use for users and improve the ease of use and popularity of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a flow chart of a coal mining data management system based on video artificial intelligence provided by an embodiment of the present invention.

[0017] Figure 2 It is a schematic diagram of a coal mining data management method based on video artificial intelligence provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in a form and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0019] like Figure 1 As shown, an embodiment of the present invention proposes a coal mining data management system based on video artificial intelligence, the system comprising: Data acquisition module 11, data access and storage module, used to access video data of various scenes in the coal mine, including data acquisition of equipment on the well, and store it in the database; The recognition module 12 extracts features and calculates trajectories of devices and personnel in the form of task slots according to different scene videos to obtain recognition results; The analysis module 13 is used to combine the recognition results with AR virtual technology enhanced visualization technology, perform fusion analysis on the recognition results, equipment condition data and video stream, and restore virtual reality through artificial intelligence recognition and AR virtual technology enhanced visualization to realize artificial intelligence intelligent scene analysis; The information selection module 14 selects feature data according to the results of the artificial intelligence intelligent scene recognition analysis, and performs multi-level statistics, analysis and prediction on information of different dimensions to obtain the selected feature scene data; The integrated management module 15 is used to integrate video surveillance, feature scene data, and character behavior analysis, and to achieve resource sharing and information exchange through unified coordination of the upper-level integrated security management platform; The control linkage protection module 16 is used to perform various interlocking protection controls at the equipment level and information level based on resource sharing and information exchange, and realize intelligent linkage of information, including early warning, recording, remote warning and linkage control; The user interface module 17 is used to provide an operation interface, including video viewing, warning information display, system settings, production data visualization, and supports multi-platform access.

[0020] In the embodiment of the present invention, efficient access to video data of various scenes in coal mines is achieved to ensure the integrity and real-time nature of data. The data collection capability of equipment on the well helps to fully understand the operating status of coal mines. The database storage function ensures the security and traceability of data. Intelligent processing of videos in the form of task slots improves the recognition accuracy of equipment and personnel. Combining AR virtual technology to enhance visualization technology makes the recognition results more intuitive and easy to understand, and the fusion analysis function helps to discover potential safety hazards and operational efficiency problems. The combination of artificial intelligence and AR virtual technology realizes the restoration of virtual reality, improves the intelligence level of scene analysis, can intelligently filter feature data, reduce manual intervention, and improve work efficiency. The integration of video monitoring and data analysis is realized, and the integrity of management is improved. The unified coordination of the upper-level comprehensive safety management platform ensures resource sharing and smooth information flow. The information management collaborative combat capability is improved, and the comprehensive safety guarantee of coal mines is strengthened. Based on resource sharing and information exchange, interlocking protection control at the equipment level and information level is realized. The intelligent linkage mechanism helps to quickly respond to abnormal situations and reduce accident losses. The early warning, recording, and remote warning functions improve the safety early warning and emergency handling capabilities of coal mines. The video viewing and early warning information display functions enhance the user's real-time perception of the status of coal mines.

[0021] In a preferred embodiment of the present invention, module 11 accesses video data of various scenes in the coal mine, including data collection of equipment above the mine, and stores it in a database, including: Deploy intelligent monitoring equipment in various key scenarios of coal mines, including fully mechanized mining, fully mechanized excavation working faces, key transportation nodes, pump rooms and substation locations; Connect the deployed intelligent monitoring equipment to the network system of the coal mine, develop a data collection interface, and connect the interface with the data management system of the coal mine to obtain a data collection interface; Through the data acquisition interface, the data sent by the intelligent monitoring equipment is collected in real time, including video images, personnel location, personnel behavior, equipment operating conditions, and safety environment factors; Store the data in the database to achieve data storage.

[0022] In the embodiment of the present invention, intelligent monitoring equipment is deployed in key scenes of coal mines to ensure real-time monitoring without blind spots and in all directions, greatly improving the coverage of safety monitoring. Specific monitoring of important areas of comprehensive mining, comprehensive excavation working faces, and key transportation nodes helps to timely discover and deal with potential safety risks. Monitoring of key locations of pump rooms and substations ensures the stable operation of coal mine infrastructure. Connecting intelligent monitoring equipment to the coal mine network system achieves smooth data transmission and reduces the possibility of data transmission delay and loss. The developed data acquisition interface is connected to the coal mine data management system to ensure data compatibility and consistency. The standardized interface design reduces the complexity of system maintenance and improves the scalability of the system. The data acquisition interface collects data sent by the intelligent monitoring equipment in real time to ensure the timeliness and accuracy of the data. The collected video images, personnel locations, and personnel behavior data provide a rich source of information for the safety management of coal mines, which helps to achieve refined management. Real-time monitoring of equipment operating conditions and safety environment factors helps to prevent equipment failures and safety accidents. The collected data is stored in the database to ensure the security and long-term preservation of the data.

[0023] In a specific embodiment of the present invention, the specific implementation process includes: Conduct a comprehensive assessment of the coal mine and determine the key scenes that need to be monitored, such as fully mechanized mining, fully mechanized excavation working faces, key transportation nodes, pump rooms, and substation locations. Select intelligent monitoring equipment based on the specific needs of each key scene. Install the selected intelligent monitoring equipment in the location of the key scene to ensure that it can fully cover and capture the required information. After installation, debug the equipment.

[0024] Through the laying of network cables, configuration of network equipment and implementation of network security measures, the deployed intelligent monitoring equipment is connected to the network system of the coal mine. According to the output format of the intelligent monitoring equipment and the requirements of the coal mine data management system, a data acquisition interface is developed.

[0025] Through the developed data acquisition interface, the data sent by the intelligent monitoring equipment is collected in real time. These data include high-definition video images, personnel location information, personnel behavior recognition results, equipment operating conditions and safety environment factors, and necessary pre-processing operations are performed on the collected raw data.

[0026] Design a reasonable database structure based on the type and scale of the collected data. Store the preprocessed data in the data table according to the designed database structure.

[0027] In a preferred embodiment of the present invention, module 12 extracts features and calculates trajectories of devices and personnel in the form of task slots according to different scene videos to obtain recognition results, including: According to the classification of video data, different task slots are created using artificial intelligence technology, and the classified video data are assigned to the corresponding task slots; In each task slot, the features of the equipment and personnel are extracted from the video data and the motion trajectory is calculated by using the feature extraction algorithm and the trajectory calculation algorithm to obtain the results of feature extraction and trajectory calculation; Based on the results of feature extraction and trajectory calculation, the identification results of equipment and personnel are generated, including the operating status of the equipment, fault warning, and the location and behavior information of the personnel.

[0028] In an embodiment of the present invention, video data is classified by artificial intelligence technology, which can efficiently distinguish videos of different scenes and improve the pertinence of data processing. Different task slots are created to process the classified video data respectively, so that parallel processing of tasks is realized and the overall processing efficiency is improved. The design of classification and task slots enables the system to more flexibly respond to the complex and changeable monitoring needs in coal mines. The application of feature extraction algorithms in each task slot can accurately extract the key features of equipment and personnel from video data. The use of trajectory calculation algorithms can accurately track the movement trajectory of equipment and personnel, which helps to discover abnormal behavior and predict potential risks. The recognition results are generated based on the results of feature extraction and trajectory calculation, which can comprehensively reflect the real-time status of equipment and personnel, including operating status, fault warning, location and behavior information.

[0029] In a specific embodiment of the present invention, the specific implementation process includes: The video data obtained from module 11 is classified in detail. The classification basis may include scene type (such as comprehensive mining face, transportation node), monitoring object (equipment, personnel) and video quality factors; using advanced video analysis technology, such as deep learning algorithm, it can identify key features in the video, assign the video to categories, and automatically classify the video content. Based on the classification results, different task slots are created using artificial intelligence technology (such as machine learning or deep learning framework). Each task slot is specifically designed to process a certain type of video data. The design of the task slot needs to consider the reasonable allocation of computing resources, including allocating appropriate computing resources (such as CPU, GPU or TPU) to each task slot.

[0030] In each task slot, specific feature extraction algorithms are used to process video data, including convolutional neural networks in deep learning, to identify key features in images. For equipment, features include shape, color, and texture; for people, features include facial features, clothing color, and movement patterns. Advanced computer vision techniques, such as optical flow and background subtraction, are used to track moving objects (equipment and people) in the video. The motion trajectory of the object is calculated by comparing and analyzing consecutive frames.

[0031] Based on the extracted features and calculated trajectories, a random forest classifier is used to identify the specific information of equipment and personnel. For equipment, its operating status can be identified, such as normal operation, shutdown, and failure; for personnel, its location and behavior can be identified, and the identification results can be displayed in a visual form, including the equipment's operating status diagram, fault warning prompts, and personnel's location and behavior reports.

[0032] In a preferred embodiment of the present invention, in each task slot, a feature extraction algorithm and a trajectory calculation algorithm are used to extract the features of the equipment and personnel from the video data and calculate the motion trajectory to obtain the results of feature extraction and trajectory calculation, including: In each task slot, the classified video data is subjected to feature extraction through a convolutional neural network to obtain key features. The key feature data includes the appearance features, operating status, and behavioral features of the equipment. Obtain the continuous time point data of the feature from the key feature, and pass Calculate the movement trajectory of equipment and personnel, where Pt It's at the time t The predicted location of is the known position at the previous time point t−1, It's at the time t−1 speed, It's at the time t−1 The acceleration of From the time point t−1To the time point t The time interval K is the correction factor for adjusting the trajectory by external forces and noise, It's at the time t′ The external forces acting on the object, Indicates at a point in time t′ Noise or measurement error, From the time point t−1 arrive t The cumulative effect of external forces and noise on the position of an object.

[0033] In an embodiment of the present invention, the convolutional neural network is used to extract features from video data, which can efficiently capture the key features of equipment and personnel, such as appearance, operating status and behavioral characteristics. The deep structure of the convolutional neural network helps to separate target features from complex video backgrounds, improving the accuracy and robustness of feature extraction. The motion trajectory of equipment and personnel is calculated by feature data at continuous time points, which achieves a detailed characterization of the dynamic behavior of the target. The introduced predicted position calculation formula takes into account the position, speed, acceleration, and external forces and noise at the previous time point, making the trajectory calculation more accurate and reliable. Trajectory calculation not only helps to track targets in real time, but also provides key inputs for abnormal behavior detection, risk assessment and early warning systems. Adjusting the correction coefficient of external forces and noise on the trajectory and considering the cumulative effects of external forces and noise enhances the adaptability of the trajectory calculation model to actual conditions and improves the system's anti-interference ability and prediction accuracy.

[0034] In a specific embodiment of the present invention, the specific implementation process includes: Receive video data from surveillance cameras or other video sources. The video data contains multiple scenes or objects. Perform preliminary classification and screening of the video data for subsequent processing. For example, different devices and personnel can be distinguished through image recognition technology. Convert the video data into a format suitable for convolutional neural network processing, such as converting video frames into images, and perform resizing and normalization.

[0035] Use pre-trained or custom convolutional neural network models to extract features from classified video data. Convolutional neural networks can automatically learn and extract key features in images, such as the appearance features of equipment (such as shape, color, texture), operating status (such as switch status, rotation speed), and behavioral features of people (such as walking speed, posture, expression). Select key features related to the task from the output of the convolutional neural network and optimize these features through principal component analysis algorithms or feature fusion techniques.

[0036] The position, velocity, and acceleration information of the equipment and personnel at continuous time points are extracted from the key features, and the given physical formula is used to calculate the motion trajectory of the equipment and personnel. The formula takes into account the position, velocity, and acceleration at the previous time point, as well as the influence of external forces and noise on the trajectory from the previous time point to the current time point. The predicted position at the current time point is calculated according to the formula, and the cumulative influence of external forces (such as wind force, friction) and noise (such as measurement error, sensor noise) on the trajectory is processed through the integral term. The adjustment coefficient K is used to control the degree of correction of these external factors on the trajectory calculation. The calculated trajectory is smoothed by a trajectory optimization algorithm (such as the least squares method) to eliminate possible noise and outliers. The system outputs the results of feature extraction and trajectory calculation, including the key features and motion trajectory information of the equipment and personnel. With the continuous addition of new data, the convolutional neural network model and trajectory calculation algorithm are continuously trained and optimized to improve the accuracy of feature extraction and trajectory calculation.

[0037] In a preferred embodiment of the present invention, module 13 combines the recognition result with AR virtual technology enhanced visualization technology, performs fusion analysis on the recognition result, equipment condition data and video stream, and restores virtual reality through artificial intelligence recognition and AR virtual technology enhanced visualization to realize intelligent scene analysis, including: Use artificial intelligence technology to perform scene analysis and behavior prediction on the recognition results to obtain scene understanding results; Based on the recognition and scene understanding results, create virtual models, including 3D representations of devices, objects, and environments; Map the equipment condition data and recognition results to the virtual model, and perform fusion analysis on the AI ​​recognition results, equipment condition data, and video streams to obtain fused data; The fused data is enhanced and displayed in the real scene through AR virtual technology, restoring a virtual scene that is highly integrated with reality; According to real-time data changes, the content and status in the AR virtual technology scene are dynamically updated to achieve intelligent scene analysis.

[0038] In the embodiment of the present invention, the recognition results are deeply analyzed by artificial intelligence technology, which can accurately understand the dynamic situation and behavior patterns in various scenes of coal mines. The behavior prediction function helps to predict potential safety risks or operational efficiency problems in advance and provide timely decision support for managers. According to the recognition results and scene understanding, highly realistic 3D virtual models of equipment, objects and environments are created. The virtual model provides an intuitive and easy-to-understand visual interface for coal mine managers to better understand the on-site situation. The 3D representation method enhances the expressiveness and interactivity of information and improves management efficiency. The equipment condition data, recognition results and virtual models are fused and analyzed to achieve the integration and association of multidimensional data. This fusion analysis helps to discover the potential connections and laws between data and provide a comprehensive perspective for the safe production and operation management of coal mines. The fused data is more valuable for reference and can support more complex decision-making and analysis needs. Through AR virtual technology, the fused data is superimposed in the real scene, realizing the organic combination of virtual and reality. This enhanced display method enables managers to experience the on-site conditions of coal mines in person, improving the perception and understanding of information. The application of AR virtual technology also improves the interactivity and immersion of the monitoring system and optimizes the user experience. The content and status of the AR virtual technology scene are dynamically updated according to real-time data changes, maintaining the authenticity and timeliness of the scene. The update mechanism enables the monitoring system to reflect the latest situation of the coal mine site in real time, providing strong support for rapid response and decision-making. The realization of the intelligent scene analysis function further improves the intelligence level of the coal mine monitoring and management system.

[0039] In a specific embodiment of the present invention, the specific implementation process includes: Receive recognition results from the front end or other modules, including information on device type, personnel behavior, and object location, and use deep learning and machine learning artificial intelligence technologies to further analyze the recognition results to understand the contextual information in the scene. This includes identifying the interaction relationship between devices, personnel behavior patterns, and object movement trajectories, and using prediction models (such as neural networks) to predict the future behavior of devices and personnel based on historical data and current scene information, and obtain scene understanding results, including device layout, personnel distribution, behavior patterns, and predicted future behaviors in the scene.

[0040] Based on the recognition results and scene understanding results, use 3D modeling software or tools to create 3D representations of devices, objects and environments, and optimize the created 3D models to improve rendering speed and visual effects, including reducing the number of polygons of the model and using texture mapping technology.

[0041] Match the data with the position of the actual equipment or object in 3D space, and use artificial intelligence algorithms to fuse and analyze the artificial intelligence recognition results, equipment condition data, and video streams. This includes data association, feature extraction, and pattern recognition steps to extract information useful for intelligent scene analysis. Map equipment condition data (such as temperature, pressure, vibration) and recognition results onto the virtual model.

[0042] Use AR virtual technology to enhance the display of fused data in real scenes, and design AR virtual technology interfaces to intuitively and clearly display fused data, including using color, shape, and animation visual elements to distinguish different data types and information. Design user interaction methods to interact with virtual information in AR virtual technology scenes by clicking, dragging, and zooming. Output the results of AR virtual technology enhanced display, and use AR virtual technology equipment to see virtual scenes that integrate equipment conditions, personnel behavior, and object location information.

[0043] Get new data from sensors and camera data sources in real time. Update the new data to the virtual model and re-analyze the fusion. Dynamically update the content and status of the AR virtual technology scene based on the updated data, including changing the color, shape, and position of virtual objects. Use artificial intelligence algorithms to perform intelligent analysis on the updated AR virtual technology scene to extract useful information or discover potential problems. Get dynamically updated AR virtual technology scenes and intelligent analysis results.

[0044] In a preferred embodiment of the present invention, module 14 selects feature data based on the results of artificial intelligence intelligent scene recognition analysis, and performs multi-level statistics, analysis and prediction on information of different dimensions to obtain selected feature scene data, including: The results of AI intelligent scene recognition analysis were evaluated using principal component analysis for dimensionality reduction to select feature data; Based on the selected characteristic data, multi-level statistical analysis is performed, including data distribution, correlation analysis, and cluster analysis; Based on the results of statistical analysis, a prediction model is constructed to predict the scenario data or behavior to obtain the selected feature scenario data.

[0045] In an embodiment of the present invention, the results of artificial intelligence intelligent scene recognition analysis are evaluated and dimensionally simplified, which can effectively evaluate and simplify the data dimensions of artificial intelligence intelligent scene recognition results, remove redundant information, and highlight key features. Through principal component analysis, the selected feature data is more representative and discriminative. Dimensional simplification also reduces the complexity of data processing and storage, and improves overall operating efficiency. Multi-level statistical analysis of the selected feature data, including distribution status, correlation analysis and cluster analysis, can fully reveal the inherent laws and potential connections of the data. The statistical analysis results provide coal mine managers with in-depth data insights, which help to better understand the scene dynamics and behavior patterns. Correlation analysis and cluster analysis can also help identify abnormal data and potential risk points, providing strong guarantees for the safe production of coal mines. Based on the results of statistical analysis, a prediction model is constructed, which can accurately predict scene data or behavior trends and provide forward-looking decision support for the operation and management of coal mines. Through the feature scene data output by the prediction model, managers can adjust production plans and resource allocation in a timely manner, optimize operation strategies, and improve the production efficiency and economic benefits of coal mines. The realization of the prediction function also improves the intelligence and automation level of coal mine monitoring and management systems, and reduces the dependence on human intervention.

[0046] In a specific embodiment of the present invention, the specific implementation process includes: Receive the results of AI intelligent scene recognition analysis, which contain a large number of features, such as equipment status, personnel behavior, and environmental parameters. Remove invalid or missing data to ensure the integrity and accuracy of the input data. Standardize the data to eliminate the dimensional differences between different features, obtain the input data, and calculate the covariance between the features of the input data. Perform eigenvalue decomposition on the covariance to obtain the eigenvalues ​​and corresponding eigenvectors. According to the size of the eigenvalues, select the eigenvectors corresponding to the first few largest eigenvalues ​​as the principal components. These principal components represent the most important information directions in the data. Project the original data onto the selected principal components to obtain the reduced-dimensional data. These reduced-dimensional data are the selected feature data.

[0047] Draw a histogram of the feature data, estimate its probability density function, understand the distribution of the data, calculate the mean, variance, median, and quartile statistics of the feature data, and evaluate the central tendency and dispersion of the data. Calculate the correlation coefficient between the feature data, use a heat map or a correlation network diagram to visualize the correlation coefficient, and intuitively display the correlation between the features. Perform a significance test on the correlation coefficient to determine which correlations are significant, that is, not caused by random errors. Select an appropriate clustering algorithm based on the characteristics of the data, such as K-means and hierarchical clustering. Use clustering effectiveness indicators (such as the silhouette coefficient) to evaluate the quality of the clustering results. Based on the clustering results, explain the scenarios or behavior patterns represented by different clusters.

[0048] Based on the results of statistical analysis and cluster analysis, select appropriate prediction models, such as linear regression, decision tree, and neural network. Before training the model, perform feature selection and optimization to improve the prediction performance of the model. This includes using feature importance scoring and recursive feature elimination methods. Use the training data set to train the model and adjust the model parameters to minimize the prediction error. Use the cross-validation method to evaluate the generalization ability of the model and prevent overfitting. Analyze the results of model prediction and identify feature scenario data with significant predictive value.

[0049] In a preferred embodiment of the present invention, multi-level statistical analysis is performed based on the selected characteristic data, including data distribution, correlation analysis, and cluster analysis, including: According to the selected characteristic data, the mean and variance of the data are calculated to obtain the distribution state of the data; According to the mean and variance of the data, correlation analysis was performed using the Pearson correlation coefficient; According to the selected feature data, Determine the final number of clusters, apply the selected clustering algorithm to perform cluster analysis on the data to obtain the cluster analysis results, where: is the average distance between a data point and other data points in the same cluster, is the minimum average distance between a data point and points in other clusters, Yes and The larger value of It is a cluster The number of data points in is the average size of all clusters, It is a cluster The area or volume of It is a cluster i The number of data points in It is a cluster i The number of data points marked as outliers in It is a cluster i The proportion of outliers in Based on the data distribution status, data correlation analysis and data cluster analysis results, multi-level statistical analysis results can be obtained.

[0050] In an embodiment of the present invention, by calculating the mean and variance of the data, the distribution state of the selected feature data can be accurately characterized. As basic statistics, the mean and variance provide intuitive indicators for understanding the central tendency and dispersion of the data. Correlation analysis using the Pearson correlation coefficient can quantify the degree of linear correlation between feature data. This analysis helps to reveal the potential relationship between different features and discover possible influencing factors or predictors. By determining the final number of clusters and applying the selected clustering algorithm, the selected feature data can be effectively clustered. Cluster analysis helps to identify natural groupings or patterns in the data and reveal hidden structures or associations. Clustering results can be used for scene classification, anomaly detection or behavior pattern recognition to improve the intelligent level of coal mine monitoring. The distribution state, correlation analysis and cluster analysis results of the comprehensive data can obtain comprehensive and in-depth multi-level statistical analysis results, and the integrated analysis provides richer data insights.

[0051] In a specific embodiment of the present invention, the specific implementation process includes: Receive the selected feature data, for each feature, calculate the average of all its observations to get the mean of the feature, for each feature, calculate the average of the squares of the differences between all its observations and the mean to get the variance of the feature. The mean and variance of each feature are obtained to describe the distribution of the data.

[0052] Receive paired observations in feature data, use the Pearson correlation coefficient formula, substitute the input data into the formula for calculation. This formula takes into account the product sum of observations, the square of the sum of observations, and the number of observations, and is used to evaluate the degree of linear correlation between two features. According to the value of the calculated correlation coefficient, the correlation between the two features is judged. The closer the absolute value of the correlation coefficient is to 1, the stronger the linear relationship between the two features; the closer the absolute value of the correlation coefficient is to 0, the weaker the linear relationship between the two features. The correlation coefficient between each feature pair is obtained to describe the correlation between the features.

[0053] Receives feature data and data point information for distance calculation. Using the given formula, calculates the distance of each data point. The final number of clusters can be determined by observing the changing trend of the value. When the value reaches the maximum or tends to be stable, the corresponding number of clusters is the optimal number of clusters. Select a suitable clustering algorithm according to the data characteristics and requirements, such as K-means and hierarchical clustering. Apply the selected clustering algorithm to perform cluster analysis on the feature data to obtain the clustering results. The clustering results include the cluster to which each data point belongs, the number of clusters, and the characteristics.

[0054] According to the above steps, the data distribution status, correlation analysis and cluster analysis results are integrated to form the comprehensive results of multi-level statistical analysis. According to the comprehensive results, the distribution characteristics of the data, the correlation between the characteristics and the clustering pattern of the data are explained to obtain the comprehensive results of multi-level statistical analysis.

[0055] In a preferred embodiment of the present invention, module 15 integrates video surveillance, feature scene data, and character behavior analysis, and realizes resource sharing and information exchange through unified coordination of the upper-level integrated security management platform, including: Analyze and plan data and establish a comprehensive security management platform; Connect video surveillance to the integrated security management platform, associate feature scene data with video surveillance, and connect character behavior analysis with the integrated security management platform to obtain an upper-level integrated security management platform; Through the unified coordination of the upper-level comprehensive security management platform, resource sharing and information exchange can be achieved.

[0056] In the embodiment of the present invention, by deeply analyzing each data, it is possible to ensure that the planning and design of the comprehensive safety management platform meets the actual needs of coal mine monitoring and management, which helps to avoid resource waste and functional redundancy in system construction and improve overall practicality and efficiency. Video surveillance, feature scene data, and behavior analysis are connected to the comprehensive safety management platform to achieve the integration and association of various data. Association between feature scene data and video surveillance. Through the unified coordination of the upper-level comprehensive safety management platform, resource sharing and information exchange are achieved. The sharing and exchange mechanism breaks the information island and improves the utilization and accuracy of data.

[0057] In a specific embodiment of the present invention, the specific implementation process includes: Gain an in-depth understanding of each data, including real-time video surveillance, video playback, and alarm triggering functions; scene recognition and anomaly detection functions of feature scene data; behavior recognition and trajectory tracking functions of the human behavior analysis system, and clarify the functions that the comprehensive security management platform needs to achieve, such as resource integration, information sharing, intelligent early warning, and emergency response. Analyze the data type, format, and storage method generated by each data to ensure the compatibility and interoperability of the data on the platform. Determine the scope and permissions of data sharing to ensure data security and privacy protection. Design the overall architecture of the comprehensive security management platform, including hardware deployment, software module division, and network interface design. Determine the platform's operating environment and dependency conditions, such as operating system, database, and middleware. According to functional requirements and data requirements, select technology stacks and tools for development, and formulate development plans, test plans, and deployment plans.

[0058] Connect the video surveillance cameras and recorders to the integrated security management platform through the network, configure the video streaming protocol, bit rate, and frame rate parameters to ensure the smoothness and clarity of the video images, and integrate and display the real-time monitoring, video playback, and alarm triggering functions of video surveillance on the platform. Connect the scene recognition results and anomaly detection information data generated by the feature scene data to the integrated security management platform, associate the feature scene data with video surveillance, and realize the synchronous display and linkage of scene recognition results and video images. Connect the behavior recognition results and trajectory tracking information data generated by the character behavior analysis to the integrated security management platform, configure the behavior analysis algorithm and parameters, ensure the accuracy and reliability of behavior recognition, realize the connection between the behavior analysis data and the integrated security management platform, and realize the integration and display of the behavior recognition results on the platform.

[0059] Establish a unified data warehouse on the comprehensive security management platform to store various data. Through data integration technology, clean, convert, and merge various data to form a unified data format and storage method, set data sharing permissions and scope, and realize data sharing and intercommunication. Establish a unified information intercommunication mechanism on the comprehensive security management platform to realize real-time information transmission and interaction, set linkage rules, and automatically trigger the response actions of other related architecture data when the data of a certain architecture triggers an alarm or anomaly, so as to realize the collaborative work of various data architectures. Use big data analysis and artificial intelligence technology to deeply mine and analyze the data on the comprehensive security management platform, establish an intelligent early warning model, predict potential safety risks and hidden dangers based on data changes and trends, formulate emergency plans and response processes, and automatically start the emergency plan and mobilize relevant resources for emergency response when the early warning model is triggered.

[0060] In a preferred embodiment of the present invention, module 16 performs various interlocking protection controls at the equipment level and information level based on resource sharing and information exchange, and realizes intelligent linkage of information, including early warning, recording, remote warning and linkage control, including: Design interlocking protection control scheme based on resource sharing and information exchange, and perform interlocking protection control on existing equipment at the equipment level and information level to achieve interlocking protection; Comprehensive testing and verification of the interlocking protection control after implementation is carried out to achieve intelligent linkage.

[0061] In an embodiment of the present invention, a more accurate and efficient interlocking protection control scheme can be designed based on resource sharing and information exchange. The implementation of interlocking protection control at the equipment level and information level significantly improves the safety of coal mine production equipment and overall information. The interlocking protection mechanism can effectively prevent potential safety accidents, reduce production risks, and ensure the safety of personnel and equipment. The interlocking protection control after implementation is comprehensively tested and verified to ensure the stability and reliability of the interlocking protection function. Through testing and verification, potential problems can be discovered and resolved in a timely manner, further improving safety. The realization of intelligent linkage enables various data to work together, improving the overall response speed and processing efficiency.

[0062] In a specific embodiment of the present invention, the specific implementation process includes: Utilize the upper-level integrated safety management platform to integrate the data resources of video surveillance, feature scene data, and character behavior analysis, standardize the data, work closely with relevant departments of coal mine enterprises, clarify the specific needs of interlocking protection control, conduct a comprehensive assessment of the potential risks in the coal mine production process, identify key equipment and important links, and determine the focus of interlocking protection control. Design specific interlocking protection logic for key equipment in coal mine production, such as coal mining machines, conveyors, and ventilators. For example, when the coal mining machine is working, if it is detected that the conveyor stops running or the ventilator fails, the power supply of the coal mining machine is automatically cut off to prevent equipment damage or safety accidents. Design equipment fault self-checking and alarm mechanism. When the equipment fails, it can automatically send out an alarm signal and trigger the interlocking protection action. When designing the interlocking protection logic, ensure that there is no conflict or danger in the collaborative work of various data. For example, when the video surveillance detects that the smoke concentration in the mine exceeds the standard, it automatically triggers the enhanced ventilation mode of ventilation and sends early warning information to relevant personnel at the same time. Design communication protocols and data exchange standards to ensure that various data can share data and information in real time and realize linkage control. Use big data analysis technology to analyze and process the collected data to predict potential safety risks and failure modes. Design an early warning mechanism to automatically send early warning information to relevant personnel and trigger interlocking protection actions when potential risks are predicted. Design a recording function to record the triggering conditions, execution actions and results of each interlocking protection control.

[0063] Simulate the coal mine production environment, build a test platform, including key equipment, sensors, and surveillance cameras, and deploy the interlocking protection control architecture on the test platform to ensure that all equipment and systems can communicate and exchange data normally. According to the design requirements of the interlocking protection control scheme, design detailed test cases, including normal operation, abnormal operation, and boundary conditions. Test the various functions of the interlocking protection control price, including equipment fault self-detection, alarm mechanism, early warning mechanism, and recording function, to verify whether the various functions meet the design requirements and whether they can accurately and timely trigger the interlocking protection action. Test the performance of the interlocking protection control price, including response time, processing speed, and stability. Ensure that the system can maintain a stable operating state under extreme conditions of high load and low latency, and accurately trigger the interlocking protection action. Test the compatibility of the interlocking protection control price with other equipment. Analyze the data collected during the test to evaluate the performance and quality of the interlocking protection control architecture. Identify problems and defects found during the test and formulate improvement measures. Verify and optimize the interlocking protection control system based on the test results and analysis conclusions.

[0064] In a preferred embodiment of the present invention, module 17 provides an operation interface, including video viewing, warning information display, system settings, production data visualization, and supports multi-platform access.

[0065] In an embodiment of the present invention, a user can directly view real-time monitoring videos of various scenes in a coal mine through an interface, and support functions of video playback, fast forward, and slow motion; monitor various abnormal situations in the coal mine production process in real time, such as equipment failures and safety hazards, and display early warning information in a timely manner through the interface to remind users to take corresponding measures. Collect and process various data in the coal mine production process, such as output, energy consumption, and equipment operating status, and display them intuitively on the interface in the form of charts and dashboards. Through the integrated operation interface, it is convenient to view videos, receive early warning information, and perform data queries and analysis, which greatly improves management efficiency. Real-time monitoring and display of early warning information helps users to promptly discover and deal with safety hazards, effectively reducing the probability of accidents. The production data visualization function enables users to intuitively understand the production situation.

[0066] like Figure 2 As shown, a method for coal mining data management based on video artificial intelligence, the method comprising: Step 21, collecting video data of various scenes in the coal mine, and collating and storing the collected video data and equipment data in a database; Step 22, extracting features of devices and personnel in the video in the form of task slots, converting the extracted features and trajectory data into artificial intelligence scene recognition results, so as to obtain artificial intelligence scene recognition results; Step 23, based on AR virtual technology enhanced visualization technology, the artificial intelligence scene recognition results, equipment condition data and original video stream are integrated and analyzed, and the virtual reality scene is restored by combining artificial intelligence recognition with AR virtual technology to realize artificial intelligence intelligent scene analysis; Step 24, based on the analysis results of the artificial intelligence intelligent scene recognition, the key feature data is screened out, and the key feature data is subjected to multi-level statistics, analysis and prediction to obtain the selected feature scene data; Step 25, integrating video surveillance, feature scene data, and character behavior analysis, and coordinating through the upper-level integrated security management platform to achieve resource sharing and information exchange between subsystems; Step 26, based on the results of resource sharing and information exchange, perform interlocking protection control at the device level and the system level to achieve intelligent linkage between the application subsystems, including automatic generation of early warning information, record preservation, remote warning sending, and linkage control; Step 27, according to the intelligent linkage between the subsystems, an operation interface is set to realize the management of monitoring and coal mining.

[0067] In an embodiment of the present invention, by collecting video data and equipment data of each scene in the coal mine, a comprehensive and detailed database can be established to provide a rich data source for subsequent data analysis, feature extraction and scene recognition. The feature extraction of equipment and personnel in the video in the form of task slots can be efficiently converted into artificial intelligence scene recognition results, improve the accuracy and efficiency of recognition, and help to achieve a deep understanding of the coal mine scene. Combined with AR virtual technology enhanced visualization technology, it is possible to achieve fusion analysis of artificial intelligence scene recognition results, equipment condition data and original video streams, provide more intuitive and realistic virtual reality scene restoration, and help to improve the depth and breadth of intelligent scene analysis. By performing multi-level statistics, analysis and prediction on the key feature data of artificial intelligence intelligent scene recognition results, more representative feature scene data can be screened out, providing more accurate and reliable data basis for subsequent decision support. By integrating video surveillance, feature scene data, and character behavior analysis, and through the unified coordination of the upper-level comprehensive safety management platform, resource sharing and information exchange between subsystems can be achieved, and the operating efficiency and response speed of the overall system can be improved. According to the results of resource sharing and information exchange, the interlocking protection control at the equipment level and system level is executed, which can realize the intelligent linkage between various application subsystems, including the automatic generation of early warning information, record keeping, remote warning sending and linkage control, effectively improving the intelligent level of coal mine safety management. By setting up the operation interface, the monitoring and coal mining management can be realized, which can provide an intuitive and convenient management tool to help managers grasp the operation status of coal mines in real time, timely discover and deal with potential safety hazards, and ensure the safety and efficiency of coal mining. At the same time, this also improves the transparency and efficiency of management and reduces management costs.

[0068] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A coal mining data management system based on video artificial intelligence, characterized in that: The system comprises: The data acquisition module is used to access the video data of various scenes in the coal mine, including the data acquisition of equipment above the mine, and store it in the database; The recognition module extracts features and calculates trajectories of devices and personnel in the form of task slots based on videos of different scenes to obtain recognition results; The analysis module is used to combine the recognition results with AR virtual technology enhanced visualization technology, integrate the recognition results, equipment condition data and video streams for analysis, and restore virtual reality through artificial intelligence recognition and AR virtual technology enhanced visualization to achieve artificial intelligence intelligent scene analysis; The information selection module selects feature data based on the results of AI intelligent scene recognition and analysis, and implements multi-level statistics, analysis, and prediction on information of different dimensions to obtain the selected feature scene data; The comprehensive management module is used to integrate video surveillance, feature scene data, and character behavior analysis, and to achieve resource sharing and information exchange through unified coordination of the upper-level comprehensive security management platform; The control linkage protection module is used to perform various interlocking protection controls at the equipment level and information level based on resource sharing and information exchange, and realize intelligent information linkage, including early warning, recording, remote warning and linkage control; The user interface module is used to provide an operation interface, including video viewing, warning information display, system settings, production data visualization, and supports multi-platform access.

2. The coal mining data management system based on video artificial intelligence according to claim 1 is characterized in that: Access video data from various scenes in coal mines, including data collection of equipment above the mine, and store them in the database, including: Deploy intelligent monitoring equipment in various key scenarios of coal mines, including fully mechanized mining, fully mechanized excavation working faces, key transportation nodes, pump rooms and substation locations; Connect the deployed intelligent monitoring equipment to the network system of the coal mine, develop a data collection interface, and connect the interface with the data management system of the coal mine to obtain a data collection interface; Through the data acquisition interface, the data sent by the intelligent monitoring equipment is collected in real time, including video images, personnel location, personnel behavior, equipment operating conditions, and safety environment factors; Store the data in the database to achieve data storage.

3. The coal mining data management system based on video artificial intelligence according to claim 1 is characterized in that: According to different scene videos, the device and personnel are featured and tracked in the form of task slots to obtain recognition results, including: According to the classification of video data, different task slots are created using artificial intelligence technology, and the classified video data are assigned to the corresponding task slots; In each task slot, the features of the equipment and personnel are extracted from the video data and the motion trajectory is calculated by using the feature extraction algorithm and the trajectory calculation algorithm to obtain the results of feature extraction and trajectory calculation; Based on the results of feature extraction and trajectory calculation, the identification results of equipment and personnel are generated, including the operating status of the equipment, fault warning, and the location and behavior information of the personnel.

4. The coal mining data management system based on video artificial intelligence according to claim 1 is characterized in that: In each task slot, the feature extraction algorithm and trajectory calculation algorithm are used to extract the features of the equipment and personnel from the video data and calculate the motion trajectory to obtain the results of feature extraction and trajectory calculation, including: In each task slot, the classified video data is subjected to feature extraction through a convolutional neural network to obtain key features. The key feature data includes the appearance features, operating status, and behavioral features of the equipment. Obtain the continuous time point data of the feature from the key feature, and pass Calculate the movement trajectory of equipment and personnel, where Pt It's at the time t The predicted location of At the previous time t−1 The known location of It's at the time t−1 speed, It's at the time t−1 The acceleration of From the time point t−1 To the time point t The time interval K is the correction factor for adjusting the trajectory by external forces and noise, It's at the time t′ The external forces acting on the object, Indicates at a point in time t′ Noise or measurement error, From the time point t−1 arrive t The cumulative effect of external forces and noise on the position of an object.

5. The coal mining data management system based on video artificial intelligence according to claim 1 is characterized in that: Combine the recognition results with AR virtual technology to enhance visualization technology, integrate the recognition results, equipment condition data and video streams for analysis, and restore virtual reality through artificial intelligence recognition and AR virtual technology to enhance visualization and achieve intelligent scene analysis, including: Use artificial intelligence technology to perform scene analysis and behavior prediction on the recognition results to obtain scene understanding results; Based on the recognition and scene understanding results, create virtual models, including 3D representations of devices, objects, and environments; Map the equipment condition data and recognition results to the virtual model, and perform fusion analysis on the AI ​​recognition results, equipment condition data, and video streams to obtain fused data; The fused data is enhanced and displayed in the real scene through AR virtual technology, restoring a virtual scene that is highly integrated with reality; According to real-time data changes, the content and status in the AR virtual technology scene are dynamically updated to achieve intelligent scene analysis.

6. The coal mining data management system based on video artificial intelligence according to claim 1 is characterized in that: According to the results of AI intelligent scene recognition and analysis, feature data is selected, and multi-level statistics, analysis and prediction are performed on information of different dimensions to obtain the selected feature scene data, including: Evaluate the results of AI intelligent scene recognition analysis to perform dimensionality reduction to select feature data; Based on the selected characteristic data, multi-level statistical analysis is performed, including data distribution, correlation analysis, and cluster analysis; Based on the results of statistical analysis, a prediction model is constructed to predict the scenario data or behavior to obtain the selected feature scenario data.

7. The coal mining data management system based on video artificial intelligence according to claim 1 is characterized in that: Based on the selected characteristic data, multi-level statistical analysis is performed, including data distribution, correlation analysis, and cluster analysis, including: According to the selected characteristic data, the mean and variance of the data are calculated to obtain the distribution state of the data; According to the mean and variance of the data, correlation analysis was performed using the Pearson correlation coefficient; According to the selected feature data, Determine the final number of clusters, apply the selected clustering algorithm to perform cluster analysis on the data to obtain the cluster analysis results, where: is the average distance between a data point and other data points in the same cluster, is the minimum average distance between a data point and points in other clusters, Yes and The larger value of It is a cluster The number of data points in is the average size of all clusters, It is a cluster The area or volume of It is a cluster i The number of data points in It is a cluster i The number of data points marked as outliers in It is a cluster i The proportion of outliers in Based on the data distribution status, data correlation analysis and data cluster analysis results, multi-level statistical analysis results can be obtained.

8. The coal mining data management system based on video artificial intelligence according to claim 1 is characterized in that: Integrate video surveillance, feature scene data, and character behavior analysis, and achieve resource sharing and information exchange through unified coordination of the upper-level integrated security management platform, including: Analyze and plan data and establish a comprehensive security management platform; Connect video surveillance to the integrated security management platform, associate feature scene data with video surveillance, and connect character behavior analysis with the integrated security management platform to obtain an upper-level integrated security management platform; Through the unified coordination of the upper-level comprehensive security management platform, resource sharing and information exchange can be achieved.

9. The coal mining data management system based on video artificial intelligence according to claim 1 is characterized in that: Based on resource sharing and information exchange, various interlocking protection controls at the equipment level and information level are carried out to realize intelligent information linkage, including early warning, recording, remote warning and linkage control, including: Design interlocking protection control scheme based on resource sharing and information exchange, and perform interlocking protection control on existing equipment at the equipment level and information level to achieve interlocking protection; Comprehensive testing and verification of the interlocking protection control after implementation is carried out to achieve intelligent linkage.

10. A method for coal mining data management based on video artificial intelligence, characterized in that: Applied to the system according to any one of claims 1 to 9, the method comprises: Collect video data of various scenes in coal mines, organize and store the collected video data and equipment data in the database; The features of the devices and people in the video are extracted in the form of task slots, and the extracted features and trajectory data are converted into artificial intelligence scene recognition results to obtain artificial intelligence scene recognition results; Based on AR virtual technology enhanced visualization technology, the artificial intelligence scene recognition results, equipment condition data and original video stream are integrated and analyzed, and the virtual reality scene is restored through the combination of artificial intelligence recognition and AR virtual technology to realize artificial intelligence intelligent scene analysis; According to the analysis results of artificial intelligence intelligent scene recognition, key feature data are screened out, and multi-level statistics, analysis and prediction are performed on the key feature data to obtain the selected feature scene data; Integrate video surveillance, feature scene data, and character behavior analysis, and coordinate through the upper-level integrated security management platform to achieve resource sharing and information exchange between subsystems; According to the results of resource sharing and information exchange, interlocking protection control at the equipment level and system level is performed to realize intelligent linkage between various application subsystems, including automatic generation of early warning information, record preservation, remote warning sending and linkage control; According to the intelligent linkage between subsystems, an operation interface is set up to realize the management of monitoring and coal mining.

Images