Visualization method and system for coal mine information

Abstract

The invention relates to the field of information visualization. The coal mine information visualization method comprises the following steps: acquiring an omnibearing mining area monitoring video; cache time delay optimization and worker visual calibration are carried out, and a plurality of worker position bounding boxes are extracted; mining area key reference point identification is carried out based on the omnibearing mining area monitoring video, three-dimensional structure modeling is carried out, and a mining area three-dimensional structure model is constructed; carrying out real-time dynamic tracking according to the plurality of worker position bounding boxes, carrying out real-time movement mapping on the mining area three-dimensional structure model, and constructing a real-time mining area twinborn model; and acquiring multi-dimensional environment monitoring parameters of the mining area, carrying out poisonous gas concentration fluctuation fitting and hazardous gas area positioning, and marking the hazardous gas area. According to the invention, real-time and dynamic coal mine information visualization is realized, and the operation safety of a mining area is improved.

Description

Technical Field

[0001] The present invention relates to the field of information visualization, and particularly to a visualization method and system for coal mine information. Background Art

[0002] With the continuous development of information technology and digital technology, the digital transformation of the coal mine industry in the fields of production, management, and safety is gradually advancing. As a high-risk and high-intensity working environment, coal mine safety production and operation efficiency are the core issues of concern in this industry. Traditional coal mine information management mostly relies on manual records, static drawings, or two-dimensional maps. The transmission and processing of information often have problems such as poor timeliness and unclear visualization, resulting in the inability to quickly and accurately obtain real-time mine information in a complex mine operation environment, increasing safety hazards and operation risks.

[0003] With the increase in coal mine mining depth and complexity, the traditional coal mine management mode and information monitoring means are difficult to meet the new requirements. The complex geological conditions, frequent environmental changes, unstable equipment operation status, and extensive personnel operation distribution inside the mine make higher requirements for the real-time monitoring and visualization of coal mine information. Especially in the event of sudden disaster accidents, how to obtain comprehensive and accurate mine information in the shortest time for rapid decision-making and emergency response has become the key to coal mine safety production management.

[0004] Most traditional coal mine information monitoring systems rely on two-dimensional floor plans or static data displays, which cannot effectively display the internal spatial layout of the mine, complex equipment operation status, and real-time operation conditions. This method often has limitations such as lagging information updates, difficulty in intuitively presenting spatial relationships, and difficulty in performing three-dimensional spatial interaction analysis, and it is difficult to provide effective decision-making support. With the maturity of Internet of Things, big data, and cloud computing technologies, real-time data collection and analysis have become important means for coal mine safety production management. However, relying solely on data collection itself is not enough to improve decision-making efficiency. How to transform these real-time data into intuitive and easy-to-understand three-dimensional visualization information has become a new challenge. Summary of the Invention

[0005] To solve the above technical problems, the present invention proposes a visualization method and system for coal mine information to solve at least one of the above technical problems.

[0006] To achieve the above object, the present invention provides a visualization method for coal mine information, including the following steps: Step S1: Obtain all-round mine area monitoring videos; perform cache latency optimization and worker vision calibration, and extract multiple worker position bounding boxes; Step S2: Identify key reference points in the mining area based on the all-round mining area monitoring video, perform 3D structure modeling, and construct a 3D structure model of the mining area; Step S3: Perform real-time dynamic tracking based on the position bounding boxes of multiple workers, and perform instant movement mapping on the 3D structure model of the mining area to construct a real-time mining area twin model; Step S4: Obtain multi-dimensional environmental monitoring parameters of the mining area, perform fitting of toxic gas concentration fluctuations and localization of dangerous gas areas, and mark the dangerous gas areas; Step S5: Infer dangerous terrains for the real-time mining area twin model, and then perform high-light visualization rendering based on the dangerous gas areas to obtain a warning map of dangerous areas; Step S6: Make real-time early warning decisions based on the warning map of dangerous areas and perform dynamic holographic visualization to construct a holographic visualization model of mining area information.

[0007] Through video surveillance and computer vision technologies, the present invention can accurately calibrate the positions of each worker, helping management personnel to grasp the distribution of workers in the mining area in real time. It reduces video stream latency, ensures data processing speed and real-time response, and provides guarantee for subsequent dynamic modeling and real-time feedback. Through accurate worker position calibration, the dynamic behaviors of workers can be monitored in real time, reducing the possibility of workers entering dangerous areas and enhancing the safety of the mining area. By identifying key reference points in the mining area, the accuracy of 3D modeling is ensured. Through the 3D model, mining area managers can comprehensively understand the spatial layout of the mining area. The 3D structure model helps to visually display the complex environment of the mining area, enhancing the spatial awareness of workers and management personnel. It provides a solid foundation for subsequent worker position calibration and warning of dangerous areas. Through dynamic tracking, the system can update the position changes of each worker in real time, ensuring that mining area management personnel accurately understand the distribution of workers. The 3D model of the mining area can be adjusted in real time according to the movement of workers, constructing a dynamic mining area twin model and providing instant data support for management decisions. Through real-time dynamic mapping and tracking, management personnel can timely detect abnormal behaviors of workers and prevent workers from entering dangerous areas. Through real-time acquisition of multi-dimensional environmental data, the system can monitor the changes in the mining area environment in real time, ensuring that managers can understand the gas conditions in the mining area at the first time. By performing fluctuation fitting on the concentration of toxic gases, the potential changes in the concentration of dangerous gases can be effectively predicted, marking dangerous areas in time and improving the predictability of safety management. By combining gas concentration and terrain features, a comprehensive warning map of dangerous areas is generated, comprehensively reflecting the risk points in the mining area. Highlight rendering and visual presentation of dangerous areas can intuitively remind managers and workers of dangerous positions, helping them to make timely responses. The visual warning map can help decision-makers and workers quickly locate dangerous areas, thus improving the response efficiency and safety guarantee ability. Real-time identification of dangerous situations and giving decision-making suggestions reduce the delay of human decision-making and ensure the safe operation of the mining area. Through dynamic holographic visualization technology, management personnel can view the real-time information of the mining area in a virtual environment, visually monitoring the operation status of the mining area and greatly improving the decision-making efficiency. Through the holographic model, workers can view their safety status and routes through augmented reality devices, thus enhancing the safety of workers.

[0008] In this specification, a visualization system for coal mine information is provided, which is used to execute the visualization method of coal mine information as described above, including: A worker vision calibration module, which is used to obtain a full-range mining area monitoring video; and perform cache latency optimization and worker vision calibration to extract multiple worker position bounding boxes; A 3D structure module, which is used to identify key reference points in the mining area based on the full-range mining area monitoring video, and perform 3D structure modeling to construct a 3D structure model of the mining area; A twin model module for real-time dynamic tracking based on multiple worker position bounding boxes, and for instantaneously moving and mapping a three-dimensional structure model of a mining area to construct a real-time twin model of the mining area; A toxic gas sensing module for obtaining multi-dimensional environmental monitoring parameters of a mining area, and for performing fitting of toxic gas concentration fluctuations and positioning of hazardous gas areas to mark hazardous gas areas; A hazardous area rendering module for inferring hazardous terrains of a real-time twin model of a mining area, and then performing highlighted visualization rendering based on the hazardous gas areas to obtain a warning map of hazardous areas; A real-time warning module for making real-time warning decisions based on the warning map of hazardous areas and for performing dynamic holographic visualization to construct a holographic visualization model of mining area information.

[0009] The present invention accurately identifies the positions of workers to ensure the monitoring of each worker, enabling quick understanding of their specific positions in case of emergencies and reducing the occurrence of work-related injuries. By optimizing video processing cache and latency issues, the bounding boxes of workers' positions can be updated in real time, improving the system's response speed, reducing latency, and ensuring the smoothness of dynamic tracking. Through 3D modeling, the spatial structure of the mining area can be visually displayed, including complex underground passages, mine facilities, etc., helping management personnel understand the layout of the mining area. The 3D model of the mining area can serve as the basis for future design and planning, improving the efficiency of resource management and optimizing the layout of mining area facilities. By real-time modeling of the 3D structure of the mining area, management personnel can detect structural problems (such as the risk of tunnel collapse) and perform maintenance or repair in a timely manner. The twin model can update the positions of workers in real time and reflect them in the 3D model, providing accurate real-time data for mining area management personnel to ensure that they can always understand the personnel dynamics in the mining area. Through real-time dynamic tracking, the twin model can help predict possible dangerous situations and take preventive measures or develop emergency response plans in advance. The real-time tracking of workers' positions helps management personnel optimize work task arrangements and avoid personnel staying in high-risk areas. Through real-time gas concentration analysis, early warnings can be issued and measures can be taken to prevent disasters such as fires or explosions caused by toxic gas leaks. The fitting of the concentration fluctuations of toxic gases can help locate risky areas, providing a basis for personnel to take shelter and evacuate. Through 3D rendering, all dangerous areas can be visually displayed in the mining area twin model, greatly improving the ability of mining area management personnel to identify potential risks. The real-time marking of dangerous areas in the mining area is helpful for the emergency management of the mining area. Staff can quickly receive instructions and take measures for emergency response. Through the rendered map of dangerous areas, workers can quickly identify and avoid dangerous areas, enhancing the safety of mining area personnel. Through the fusion analysis of real-time data, the mining area can respond quickly to potential safety accidents, issue early warnings, and reduce the probability of accidents. Dynamic holographic visualization enables mining area management personnel to more clearly see dangerous areas and the overall safety status of the mining area, helping to make quick and accurate decisions. Through the holographic visualization model of the mining area, various types of dangers occurring in the mining area can be simulated, further enhancing the safety prevention and control ability of the mining area. Description of the Drawings

[0010] Figure 1 It is a schematic flow chart of the steps of a visualization method for coal mine information according to the present invention; Figure 2 It is a schematic detailed implementation step flow chart of step S1; Figure 3 It is a schematic detailed implementation step flow chart of step S2; Figure 4 It is a schematic detailed implementation step flow chart of step S3. Detailed Implementation Manner

[0011] It should be understood that the specific embodiments described herein are for explaining the present invention only and are not used to limit the present invention.

[0012] The embodiments of the present application provide a visualization method and system for coal mine information. The execution subjects of the visualization method and system for coal mine information include but are not limited to: mechanical equipment, data processing platforms, cloud server nodes, network uploading devices, etc. that are equipped with this system and can be regarded as general computing nodes of the present application. The data processing platform includes but is not limited to at least one of: audio and image management systems, information management systems, and cloud data management systems.

[0013] Please refer to Figures 1 to 4 , the present invention provides a visualization method for coal mine information. The visualization method for coal mine information includes the following steps: Step S1: Obtain an all-round mine area monitoring video; perform cache delay optimization and worker vision calibration, and extract multiple worker position bounding boxes; Step S2: Identify key reference points in the mine area based on the all-round mine area monitoring video, and perform three-dimensional structure modeling to construct a three-dimensional structure model of the mine area; Step S3: Perform real-time dynamic tracking based on multiple worker position bounding boxes, and perform instant movement mapping on the three-dimensional structure model of the mine area to construct a real-time mine area twin model; Step S4: Obtain multi-dimensional environmental monitoring parameters of the mine area, perform fitting of toxic gas concentration fluctuations and positioning of dangerous gas areas, and mark the dangerous gas areas; Step S5: Infer dangerous terrains for the real-time mine area twin model, and then perform high-light visualization rendering based on the dangerous gas areas to obtain a warning map of dangerous areas; Step S6: Make real-time early warning decisions based on the warning map of dangerous areas and perform dynamic holographic visualization to construct a holographic visualization model of mine area information.

[0014] Through video surveillance and computer vision technologies, the present invention can accurately calibrate the positions of each worker, helping management personnel to grasp the distribution of workers in the mining area in real time. It reduces video stream latency, ensures data processing speed and real-time response, and provides guarantee for subsequent dynamic modeling and real-time feedback. Through accurate worker position calibration, the dynamic behaviors of workers can be monitored in real time, reducing the possibility of workers entering dangerous areas and enhancing the safety of the mining area. By identifying key reference points in the mining area, the accuracy of 3D modeling is ensured. Through the 3D model, mining area managers can comprehensively understand the spatial layout of the mining area. The 3D structure model helps to intuitively display the complex environment of the mining area, enhancing the spatial awareness of workers and management personnel. It provides a solid foundation for subsequent worker position calibration and dangerous area warning. Through dynamic tracking, the system can update the position changes of each worker in real time, ensuring that mining area management personnel accurately understand the distribution of workers. The 3D model of the mining area can be adjusted in real time according to the movement of workers, constructing a dynamic mining area twin model and providing instant data support for management decisions. Through real-time dynamic mapping and tracking, management personnel can timely discover abnormal behaviors of workers and prevent workers from entering dangerous areas. Through real-time acquisition of multi-dimensional environmental data, the system can monitor the changes in the mining area environment in real time, ensuring that managers can understand the gas conditions in the mining area at the first time. Fluctuation fitting is performed on the concentration of toxic gases to effectively predict potential changes in the concentration of dangerous gases, timely mark dangerous areas, and improve the predictability of safety management. By combining gas concentration and terrain features, a comprehensive warning map of dangerous areas is generated, comprehensively reflecting the risk points in the mining area. Highlight rendering and visual presentation of dangerous areas can intuitively remind managers and workers of dangerous positions, helping them to respond in a timely manner. The visual warning map can help decision-makers and workers quickly locate dangerous areas, thereby improving the response efficiency and safety guarantee ability. Real-time identification of dangerous situations and giving decision-making suggestions reduce the delay of human decision-making and ensure the safe operation of the mining area. Through dynamic holographic visualization technology, management personnel can view the real-time information of the mining area in a virtual environment, intuitively monitor the operation status of the mining area, and greatly improve the decision-making efficiency. Through the holographic model, workers can view their safety status and routes through augmented reality devices, thereby improving the safety of workers.

[0015] In an embodiment of the present invention, refer to Figure 1 , which is a schematic diagram of the step flow of a visualization method for coal mine information of the present invention. In this example, the steps of the visualization method for coal mine information include: Step S1: Obtain all-round mining area monitoring videos; perform cache latency optimization and worker vision calibration, and extract multiple worker position bounding boxes; In this embodiment, multiple high-resolution cameras are arranged within the mining area to ensure coverage of the key areas of the entire mining area. These cameras should have night vision capabilities and earthquake resistance to operate under various environmental conditions. Suppose 10 cameras are arranged in the mining area, distributed in the main operation areas, entrances and exits, and dangerous areas, ensuring that at least one camera can monitor each key area. Configure the video acquisition system and set the acquisition parameters (such as resolution, frame rate, and video encoding format). Usually, the video resolution is selected as 1080p, and the frame rate is set to 30fps to ensure that the movements of workers can be captured. Select the H.264 codec for video compression to reduce storage space while ensuring video quality. Set the storage capacity of each camera to 1TB to ensure that at least one week's worth of monitoring videos can be stored. Transmit the acquired video data to the central control system in real time, using a stable network protocol (such as RTSP or RTMP) for data transmission. Ensure the real-time and integrity of the video data. Set the network bandwidth to 10Mbps to support the simultaneous transmission of video streams from multiple cameras, ensuring that the data of each camera can be seamlessly connected to the control center. During the video acquisition and transmission process, monitor the latency performance of the monitoring system and analyze the time delay of the video stream transmission. Use network monitoring tools (such as Wireshark) to evaluate network latency and data packet loss rate. If it is found that the latency during certain periods exceeds 200 milliseconds, optimization is required to ensure the real-time nature of the monitoring. Design and implement an effective caching strategy, using a circular buffer or queue structure to store the most recent video data. This strategy can help reduce transmission latency and improve the stability of the video stream. Set the buffer size to 30 seconds of video data to prevent data loss and ensure the continuity of the data stream in case of poor network conditions. After implementing the caching strategy, continuously monitor the network latency to ensure the optimization effect. Record the latency changes before and after optimization and evaluate the effectiveness of the caching strategy. After optimization, the monitored latency drops to 150 milliseconds, meeting the requirements of real-time monitoring and ensuring that the activities of workers can be reflected in a timely manner. Prepare the calibration scenario to ensure that the cameras can accurately capture the positions of workers. Calibration markers (such as colored dots or squares) can be set to identify the positions of workers in the video. Use calibration tools within the key areas of the mining area and place multiple distinctively colored markers on the ground for subsequent algorithm recognition. Select and apply a suitable worker detection algorithm (such as real-time object detection algorithms like YOLO, SSD, etc.) to analyze each frame of the video stream and extract the bounding boxes of workers. Ensure that the algorithm can work stably under different lighting conditions. Use the YOLOv5 model and set the input resolution to 640x640 to ensure that workers can be accurately detected and bounding boxes can be generated in each frame. Record the information of the detected bounding boxes of workers' positions, including the coordinates of each worker's bounding box and the recognition confidence, etc. Ensure that all data can be stored in the database for subsequent analysis and visualization.The record format is "timestamp, worker ID, bounding box coordinates (x1, y1, x2, y2), confidence level" to ensure the integrity and traceability of the data.

[0016] Step S2: Identify key reference points in the mining area based on the omnidirectional mining area monitoring video, and perform three-dimensional structure modeling to construct a three-dimensional structure model of the mining area; In this embodiment, key frames are extracted from the all-round mining area monitoring video for subsequent reference point recognition. The extraction frequency is set to 1 frame per second to ensure the capture of dynamic changes in the mining area. Assuming the monitoring video duration is 10 minutes, a total of 600 frames are extracted. Each extracted frame is preprocessed, including denoising, contrast enhancement, etc., to improve the accuracy of subsequent feature extraction. In each frame, key reference points are identified using feature point detection algorithms (such as SIFT, ORB, or SURF). These feature points are usually objects with significant textures or shapes in the mining area, such as ore piles, equipment, and signboards. Using the ORB algorithm, 500 feature points are detected in each frame image, and the threshold is set to 10 to ensure that only the most significant feature points are retained for subsequent analysis. The coordinates and descriptors of each feature point are recorded. The feature points in each frame are matched using the FLANN (Fast Library for Approximate Nearest Neighbors) or BFMatcher algorithm to associate the same feature points in different frames. By setting a matching degree threshold (such as 0.75), high-quality matching pairs are filtered out. If the feature point A in a certain frame has a match in the next frame and the matching degree exceeds the threshold, then this feature point pair is recorded. Finally, about 350 pairs of stably matched feature points are retained to ensure the reliability of the feature points. Before performing 3D modeling, the camera needs to be calibrated to obtain the internal and external parameters of the camera. Multiple calibration images are taken using the checkerboard calibration method, and calibration is performed by calculating the camera matrix. 10 checkerboard images at different angles are taken, and the Zhang Zhengyou calibration method is used to calculate the internal parameter matrix and distortion coefficients of the camera to ensure the accuracy of 3D reconstruction. According to the feature matching results, the triangulation method is used to generate a 3D point cloud. Through the internal and external parameters of the camera, the pixel coordinates of each pair of feature points are converted into 3D world coordinates. If the pixel coordinates of feature point A are (x1, y1) and (x2, y2), its 3D coordinates (X, Y, Z) are calculated through the camera parameters, and finally a point cloud data set containing thousands of 3D points is generated. The generated point cloud data is processed, including downsampling, denoising, and point cloud registration. The Voxel Grid filter is used to downsample the point cloud to reduce the data volume and improve the calculation efficiency. The point cloud data is reduced from 100,000 points to 20,000 points. Subsequently, the ICP (Iterative Closest Point) algorithm is used to register the point clouds from different perspectives to ensure the accuracy and integrity of the point cloud data. Based on the processed point cloud data, 3D modeling software (such as Meshlab or Blender) is used to construct a 3D model. By generating a mesh and texture mapping, the point cloud data is converted into a visual 3D model.Generate an optimized point cloud data into a mesh model containing 5,000 triangles, and add texture mapping to it to ensure that the appearance of the mining area can be truly reflected during visualization.

[0017] Step S3: Perform real-time dynamic tracking based on multiple worker position bounding boxes, and perform instant movement mapping on the three-dimensional structure model of the mining area to construct a real-time mining area twin model; In this embodiment, the worker position bounding box data extracted from the previous step is updated in real time. Each bounding box should contain the position information of the worker (such as coordinates, width, and height) and the recognition confidence. The monitoring system receives the video stream and bounding box data in real time to ensure the accuracy and timeliness of the data. Assume that the monitoring system updates the position of the worker every second. The bounding box of worker A is recorded in real time as (220, 340, 50, 70), indicating that the center point coordinates are (220, 340), the width of the bounding box is 50 pixels, and the height is 70 pixels. A dynamic tracking algorithm (such as the Kalman filter or the SORT (Simple Online and Realtime Tracking) algorithm) is used to track the worker's bounding box in real time. These algorithms can handle changes in the worker's position and improve the accuracy of tracking through prediction and update steps. If the worker moves from the position (220, 340) to (225, 345) in the monitoring video, the Kalman filter can predict the next position based on the worker's speed and acceleration to ensure the continuity of tracking. The tracking results are recorded in the database in real time, including the worker's ID, timestamp, current position, and bounding box information. Ensure that all data can be accessed at any time for subsequent analysis and visualization. Ensure that the three-dimensional structure model of the mining area can be combined with the real-time position data of the workers. The model should contain information such as the terrain, equipment, and main channels of the mining area. Convert the model into a format suitable for real-time update (such as a format that can be used in virtual reality or augmented reality). The coordinate system of the mining area model is consistent with the coordinate system of the worker's position to ensure that the worker's position can be directly mapped into the three-dimensional model. According to the real-time tracked worker position, apply a mobile mapping algorithm to update the worker's position into the three-dimensional model. Linear interpolation or other interpolation methods can be used to smooth the worker's movement path to ensure the smoothness when the model is updated. If worker A moves from (220, 340) to (225, 345) within 1 second, then update its position on the model from (220, 340, 0) to (225, 345, 0) to ensure an accurate representation on the ground. Render the updated worker position in real time into the three-dimensional structure model of the mining area to ensure that the management personnel can intuitively see the real-time distribution state of the workers in the mining area. Use a graphics rendering engine (such as Unity or Unreal Engine) for visualization. Integrate the real-time position information of the workers with the three-dimensional model of the mining area to form a real-time mining area twin model. Ensure the synchronization and consistency of the data to be able to reflect the actual situation of the mining area in real time. Set an update frequency of once per second to ensure that each update can accurately reflect the position and status of the workers. In the real-time mining area twin model, provide a global view and a local view for the management personnel to conduct comprehensive monitoring. The global view shows the situation of the entire mining area, while the local view highlights the specific area where the worker is located.Display the layout of the entire mining area in the global view and highlight the area where Worker A is located in the local view to facilitate quick identification of his position.

[0018] Step S4: Obtain the multi-dimensional environmental monitoring parameters of the mining area, perform fitting of the toxic gas concentration fluctuations and locate the hazardous gas areas, and mark the hazardous gas areas; In this embodiment, a variety of sensors are deployed in the mining area, including oxygen sensors, combustible gas sensors (such as methane and carbon monoxide), toxic gas sensors (such as hydrogen sulfide), temperature and humidity sensors, and barometric pressure sensors. Ensure that the sensors cover the key areas of the mining area to comprehensively monitor the environmental conditions. Assume that 12 gas sensors, 6 oxygen sensors and other environmental parameter sensors are deployed in the mining area, and ensure that each sensor can accurately measure and reach the set sensitivity level (such as the detection sensitivity of the gas sensor is 1 ppm). Configure a data acquisition system, set the data acquisition frequency (such as once per second), and ensure that the acquired data can be transmitted to the central monitoring system in real time. Use a stable communication protocol (such as LoRa or Zigbee) for data transmission to ensure the real-time and integrity of the data. Set the acquisition period of each sensor to 1 second to ensure the data stability of the sensors under different environmental conditions (such as high humidity and high temperature), and use a data caching mechanism to prevent data loss. Extract the real-time monitored toxic gas concentration data from the database and perform data cleaning and preprocessing. Remove outliers and noise to ensure the accuracy and reliability of the data. Use statistical methods (such as Z-score) to identify and eliminate data points with abnormal concentration fluctuations to ensure the accuracy of subsequent fluctuation fitting. Select a suitable mathematical model for the fluctuation fitting of the toxic gas concentration. Commonly used models include linear regression, polynomial fitting, and time series analysis methods (such as ARIMA model), and select the model most suitable for the current data characteristics. If the extracted methane concentration data shows non-linear changes over a period of time, select a quadratic polynomial fitting model for concentration fluctuation analysis. Apply the selected fitting model to model the toxic gas concentration data and generate a concentration fluctuation curve. Evaluate the fitting effect of the model by calculating the goodness of fit (such as R² value), and adjust the model parameters to optimize the fitting result. According to the fluctuation fitting result of the toxic gas concentration, set the threshold standard for the dangerous gas concentration. Generally, if the concentration of a certain gas exceeds the safety threshold (such as the methane concentration exceeds 1.5%), it is marked as a dangerous area. Set the safety threshold of carbon monoxide to 50 ppm. If the monitored concentration exceeds this value, the area should be regarded as a high-risk area. Continuously monitor the data of each sensor and judge in real time whether there is a dangerous gas concentration exceeding the set threshold. Once it is found that the toxic gas concentration in a certain area exceeds the threshold, immediately mark the area as a dangerous area. If the carbon monoxide concentration of 70 ppm is detected in a certain monitoring area, mark the area as a "dangerous gas area" and record the relevant parameters. Mark the identified dangerous gas areas to ensure that they can be clearly displayed on the visualization platform of the mining area. Use different colors or symbols to represent different types of dangerous areas (such as red for high-risk areas and yellow for medium-risk areas).

[0019] Step S5: Perform dangerous terrain inference on the real-time mine twin model, and then perform highlighted visualization rendering based on the dangerous gas area to obtain a dangerous area warning map; In this embodiment, using the 3D model of the mining area and environmental monitoring data, terrain information is collected, including slope, undulation, and potential landslide areas, etc. These terrain features are extracted through graphic processing software to ensure accurate data. The digital elevation model (DEM) is used to generate the elevation data of the mining area, and the average slope of each area is calculated. If the slope of a certain area exceeds 30°, it is regarded as potential dangerous terrain. According to the operation requirements and safety standards of the mining area, the identification criteria for dangerous terrain are set. Areas with a slope exceeding the set threshold (such as 30°) or having obvious geological fractures, fissures and other characteristics are identified as dangerous terrain. If the calculated slope of a certain area is 35° and there are cracks on the ground surface, then this area is marked as "high-risk terrain". An algorithm (such as a rule-based inference system or a machine learning model) is used to analyze the collected terrain data to infer potential dangerous terrain areas. This algorithm can comprehensively consider various terrain features for comprehensive judgment. If an area meets both the slope and crack dangerous criteria at the same time, then this area is marked as "dangerous terrain", and its specific location and type are recorded. The dangerous terrain areas are integrated with the previously identified dangerous gas area data to form a comprehensive dangerous area dataset. Ensure that all data is convenient for subsequent visualization processing. If the area where the carbon monoxide concentration exceeds the safety threshold overlaps with the high-slope area, it is marked as a "comprehensive dangerous area". Select a suitable visualization rendering technology to highlight the dangerous areas in different colors or symbols in the 3D model of the mining area. A graphic rendering engine (such as Unity, Unreal Engine) can be used for effect display. All dangerous terrain areas are highlighted in red, and yellow indicates areas where the gas concentration exceeds the standard, ensuring that areas with different risk levels can be clearly distinguished. Configure the visualization system so that it can update and display the information of the dangerous areas in real time. When the monitoring system detects a change in the status of the dangerous area, this change is immediately reflected in the 3D model. If the gas concentration in a certain area returns to the safe level, then this area changes from red to green in the visualization model and is fed back to the management personnel in real time. According to the integrated dangerous area data, design and generate a dangerous area warning map. The map should include the overall layout of the mining area, dangerous terrain, and areas where the gas concentration exceeds the standard, and ensure that the information is clearly readable. Set the map scale to 1:1000 to ensure that the location and size of each dangerous area can be accurately marked on the map. Integrate the marking information of the dangerous areas into the warning map, and add necessary legends and explanations to facilitate users to understand the risk levels of each area. The legend indicates that red represents high-risk terrain and yellow represents areas where the gas exceeds the standard, and users can quickly identify the safety status of the current mining area according to the legend.

[0020] Step S6: Based on the warning map of the dangerous area, make real-time early warning decisions and conduct dynamic holographic visualization to construct a holographic visualization model of the mine area information.

[0021] In this embodiment, a real-time monitoring system is configured in the mining area to continuously monitor the status changes of dangerous areas. The system should be able to obtain the latest data from sensors and video surveillance, and compare this information with the warning map of dangerous areas. Set the data acquisition frequency to once per second to ensure that the system can receive the latest data from gas sensors and video surveillance in real time and update the status of dangerous areas in real time. Design a real-time warning mechanism. When it is detected that a worker enters a dangerous area or the concentration of dangerous gas exceeds the threshold, the system should immediately issue an alarm and notify the relevant personnel. A conditional trigger mechanism can be used to execute these warnings. It is set that if a certain worker enters an area marked as "high risk", the system immediately triggers an alarm, records the entry time, worker ID, and dangerous area information, and sends them to the terminal of the management personnel. Display the warning information through the interface of the monitoring system, including the specific location of the dangerous area, the identity of the affected workers, and specific safety suggestions (such as evacuation or taking protective measures). On the monitoring interface, mark the dangerous area with a red frame and display the alarm information next to it, such as "Worker A is in the dangerous area, please evacuate immediately". Select a suitable holographic visualization technology (such as augmented reality (AR) or virtual reality (VR)) to display the real-time data and dangerous area information of the mining area. Ensure that the technology can support dynamic updates and multi-user interactions. Use AR technology to overlay the three-dimensional model of the mining area with real-time monitoring data to ensure that the management personnel can intuitively see the relationship between each worker, their location, and the dangerous area. Integrate the real-time monitoring data (such as worker location, dangerous gas concentration, and dangerous area information) into the holographic visualization model. Ensure that the model can be updated in real time and reflect the status of the mining area according to the latest data changes. If the gas concentration in a certain area returns to a safe level, the system should change the color of that area from red to green in the holographic model to dynamically reflect the change in the safe status. Verify the constructed holographic visualization model to ensure that it accurately reflects the actual situation of the mining area. It can be verified by comparing with the on-site data to check whether the markings of worker locations and dangerous areas in the model are consistent. If the error between the worker location in the holographic model and the real-time monitoring data is less than 5 meters, it is recorded as passing the verification to ensure the credibility of the model. Collect the usage feedback of the management personnel on the holographic visualization model, analyze the problems in the user experience, and optimize the system. Ensure the usability and practicality of the model to better support the safety management of the mining area. If the management personnel feedback that some functions are difficult to use, adjust the interface and optimize the functions according to the feedback to ensure that users can quickly access the required information. Regularly update the base map and data of the holographic visualization model to ensure that the model always reflects the latest status of the mining area. Establish a maintenance plan to ensure the long-term stable operation of the system. Update the three-dimensional model of the mining area monthly to reflect the new equipment layout and environmental changes to ensure that the information obtained by the management personnel is always the latest.

[0022] In this embodiment, refer to Figure 2, which is a schematic diagram of the detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include: Step S11: Obtain an all-round mining area monitoring video based on multiple high-definition cameras in the mining operation area; Step S12: Optimize the cache delay of the all-round mining area monitoring video to construct a delay-optimized monitoring video; Step S13: Perform deep vision detection processing on the delay-optimized monitoring video to mark each worker node; Step S14: Calculate the real-time position of the worker nodes to generate the real-time position coordinates of each worker; Step S15: Calibrate the image bounding box of the delay-optimized monitoring video according to the real-time position coordinates of each worker, and extract multiple worker position bounding boxes.

[0023] In this embodiment, within the mining operation area, a number of high-definition cameras are reasonably arranged to ensure complete coverage of the entire operation area. Cameras with high resolution (such as 1080p or higher) should be selected to facilitate capturing clear video images. If the operation area is 1000 square meters, considering the field of view coverage, at least 6 cameras may need to be arranged at different corners and key positions to ensure effective monitoring of each worker's activity area. Configure a video acquisition server to be able to receive video streams from each camera in real time. Ensure that the system has sufficient bandwidth and storage capacity to handle high-definition video data. Set the network bandwidth of the system to at least 10 Mbps to support the concurrent transmission of multiple video streams. Use an efficient coding format (such as H.264) to optimize the transmission efficiency and storage space of video streams. To optimize the video cache latency, design an efficient video cache mechanism. This mechanism should be able to dynamically adjust the cache policy according to the network conditions and the priority of video streams to ensure real-time performance and smoothness. Set the priority rules, where the video streams in important worker areas have a higher priority than other areas to ensure that the video streams in key areas will not be interrupted when the bandwidth is insufficient. Analyze the cache latency of video streams, identify the main factors affecting latency (such as network fluctuations, storage speed, etc.), and apply optimization algorithms for adjustment. Utilize adaptive bitrate streaming (ABR) technology to dynamically adjust the resolution and frame rate of video streams according to network conditions to ensure the coherence of video streams even when the bandwidth fluctuates. After optimizing the video cache, generate a latency-optimized monitoring video to ensure a smooth visual experience in real-time monitoring. Record the latency data before and after optimization for subsequent analysis. The latency of the optimized video stream should be controlled within 100 milliseconds to ensure that the real-time dynamics of workers can be promptly reflected in the monitoring system. Utilize deep learning techniques (such as convolutional neural network CNN) to train a visual model for worker detection. This model should be able to accurately identify workers in the video and mark them. Use an annotated dataset containing multiple worker activity scenarios for training to ensure that the model can recognize the postures and actions of different workers. Input the latency-optimized monitoring video into the deep visual detection model for real-time processing. The model should be able to identify each worker and mark them in the video frame. When the model outputs the marking results, a bounding box can be drawn above each worker and the identity information of the worker (such as worker number or name) can be annotated. Record the detected worker node information (such as position coordinates and marking time) in the database for subsequent query and analysis. Ensure the integrity and real-time performance of the data. If the coordinates of a certain worker in the video frame are (300, 450), it is recorded as "worker number, timestamp, coordinates". By analyzing the worker markings in the video frame, calculate the real-time position coordinates of each worker. Considering the resolution of the video, ensure the accuracy of the coordinates is consistent with the actual scenario.If the video resolution is 1920x1080 and the coordinates of the worker in the image are (300, 450), then convert these coordinates into the actual working area coordinates and correct them in combination with the installation height and angle of the camera. According to the real-time position coordinates of each worker, draw the corresponding bounding boxes in the monitoring video, extract the position bounding boxes of multiple workers for subsequent data analysis and visualization. If the bounding box coordinates of a certain worker are from (290, 440) to (310, 460), then draw this box in the video and record the position information of the box. Integrate the extracted position bounding boxes of multiple workers to generate a comprehensive worker monitoring report, including the real-time coordinates and status information of each worker, and perform visualization in the monitoring system. Use the graphical user interface (GUI) to display the real-time positions and working statuses of each worker to form a dynamic monitoring platform for safety management and scheduling.

[0024] In this embodiment, the specific steps for caching delay optimization of the omnidirectional mining area monitoring video and constructing a delay-optimized monitoring video are as follows: Perform frame-by-frame video histogram brightness distribution recognition on the omnidirectional mining area monitoring video to obtain the brightness distribution characteristics of each frame; Calculate the global average brightness value of the omnidirectional mining area monitoring video; Detect the brightness deviation positions of the brightness distribution characteristics of each frame according to the global average brightness value, and mark the brightness deviation areas of each frame; Perform local brightness optimization on the brightness deviation areas of each frame to construct a brightness-optimized monitoring video; Perform abnormal noise point recognition on the brightness-optimized monitoring video to mark multiple video noise points; Calculate the pixel values of the video noise points; Perform adaptive filtering denoising according to the pixel values to obtain a filtered and optimized monitoring video; Calculate the decoding delay of the filtered and optimized monitoring video to identify the video decoding delay parameters; Calculate the highest bit rate of the current network state; Perform dynamic bit rate control optimization based on the highest bit rate and the video decoding delay parameters to generate a dynamically optimized bit rate; Perform caching delay optimization on the filtered and optimized monitoring video based on the dynamically optimized bit rate to construct a delay-optimized monitoring video.

[0025] In this embodiment, the omnidirectional mining area monitoring video is processed frame by frame to extract each frame of image. A video processing tool (such as OpenCV) can be used for frame extraction to ensure that each frame can be analyzed individually. The video frame extraction frequency is set to 30 frames per second. If the video duration is 10 minutes, a total of 1,800 frames of images are extracted. The brightness histogram of each frame of image is calculated to obtain the brightness distribution characteristics of each frame. The brightness histogram can reflect the number of pixels with different brightness values in the image. Usually, the image is converted to a grayscale image for analysis. If the histogram of a certain frame shows that the brightness values are between 0 and 255, and the calculation result shows that the number of pixels with a brightness value of 100 is 500 and the number of pixels with a brightness value of 150 is 300, then these characteristics are recorded for subsequent analysis. According to the brightness distribution characteristics of all frames, the global average brightness value is calculated. The global average brightness value can be obtained by adding up the brightness values of each frame and dividing by the total number of frames. If the total brightness value obtained from 1,800 frames is 150,000, the calculated global average brightness value is: Global average brightness = 150,000 / 1,800 = 83.33. The brightness distribution characteristics of each frame are compared with the global average brightness value to identify the brightness deviation area. A threshold (such as ±10) can be set to determine whether there is an obvious brightness deviation. If the average brightness of a certain frame is 75, compared with the global average brightness of 83.33, and the deviation is found to exceed 10, then this frame is marked as a brightness deviation frame. In each frame, the brightness deviation area is marked. Image processing techniques (such as contour detection) can be used to identify the specific deviation area and draw a bounding box on the image. If the deviation area is a rectangular area in the upper left corner, a red border is drawn within this area to clearly identify the brightness deviation.

[0026] For the brightness deviation area in each frame, a local brightness optimization algorithm is applied for processing. Techniques such as histogram equalization or contrast-limited adaptive histogram equalization (CLAHE) can be used to optimize the brightness. If the brightness of a certain deviation area is too low, the brightness of this area can be increased through the CLAHE technique to make it close to the global average brightness. The processed frames are recombined to generate a monitoring video with optimized brightness. Ensure that the brightness distribution of the optimized video is more uniform, improving the visualization effect. If after optimization, the overall brightness distribution of the video is more balanced, it is convenient for subsequent analysis and monitoring. Image processing techniques (such as Gaussian filtering) are used to detect noise in the brightness-optimized monitoring video to identify abnormal noise. Noise usually appears as points with a large brightness difference from the surrounding area. A brightness difference threshold is set. If the brightness of a certain pixel differs from the average value of the surrounding pixels by more than the threshold, it is marked as noise. The detected noise is marked, and its pixel value is recorded. Ensure that the problem area in the video can be quickly located. If the coordinates of a certain noise are (500, 300) and its pixel value is 255, it is recorded as "noise coordinates, pixel value".

[0027] According to the identified noise pixel values, apply an adaptive filtering algorithm (such as median filtering) for denoising. This algorithm can effectively remove isolated noise points while preserving edge information. If multiple noise points are detected in a certain area, use median filtering to adjust the pixel values in that area to make them closer to the median of the surrounding pixels. After denoising, recombine the processed frames to generate a monitored video optimized by filtering. Ensure that the noise in the video is effectively removed to improve the video quality. After filtering, the visual effect of the video is clearer, facilitating subsequent monitoring and analysis. Calculate the decoding delay of the monitored video optimized by filtering, measuring the time delay from the input of the video signal to the display output. Precise measurement can be carried out by setting timestamps. If the measured delay during the decoding process is 200 milliseconds, record this parameter for subsequent analysis. Record the calculated decoding delay parameter in the database and associate it with the video processing result for subsequent performance evaluation. The recording format is "video frame number, decoding delay time" to ensure the integrity of the data. By monitoring the network status in real time, calculate the maximum bitrate in the current environment, which can be done through a network bandwidth testing tool to ensure accurate evaluation of network performance. If the current network bandwidth test result is 10 Mbps, the maximum bitrate can be calculated to adapt to the transmission of the video stream. Record the maximum bitrate in the database for subsequent use.

[0028] Ensure that when performing dynamic bitrate control, the current network condition can be accurately referred to. Based on the maximum bitrate and decoding delay parameter of the current network status, establish a dynamic bitrate control model. This model should be able to adjust the bitrate of the video stream according to the real-time network condition to ensure smooth playback. Set an adaptive bitrate control strategy. When the network bandwidth is lower than 5 Mbps, automatically reduce the video quality to 720p. Through model calculation, generate a dynamically optimized bitrate and apply it to the video stream transmission to ensure good video quality and playback smoothness under different network conditions. If the network condition is good, set the bitrate to 5 Mbps; if the network condition is poor, reduce it to 2 Mbps. Based on the dynamically optimized bitrate, optimize the video caching delay. Ensure that during the streaming media playback process, the video quality and smoothness can be effectively balanced. When the network condition is good, set a larger cache; when the network condition is unstable, reduce the cache to improve the response speed. Integrate all the optimization steps to generate a final latency-optimized monitored video to ensure good visual effects and smoothness during video playback. The finally generated video can adaptively adjust under different network conditions to provide a stable monitoring experience.

[0029] In this embodiment, refer to Figure 3 , which is a schematic diagram of the detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of the said step S2 include: Step S21: Identify key reference points in the mining area from the latency-optimized monitoring video to obtain multiple mining area feature points; Step S22: Analyze the three-dimensional structure of the mining area for multiple mining area feature points to generate three-dimensional structure features of the mining area; Step S23: Perform spatial topology analysis of feature points based on multiple mining area feature points to obtain the spatial topology relationship of feature points; Step S24: Perform multi-level three-dimensional modeling on the three-dimensional structure features of the mining area based on the spatial topology relationship of feature points to construct a three-dimensional structure model of the mining area.

[0030] In this embodiment, key reference points are identified in the mining area other than the mining area. These reference points are usually objects with significant features in the mining area, such as ore piles, equipment, signboards, etc. The Scale-Invariant Feature Transform (SIFT) or Oriented FAST and Rotated BRIEF (ORB) algorithm is used to extract features from each frame to ensure accurate identification of different key reference points. Suppose 10 key points are detected in the video, located at different positions in the mining area. Through a feature point matching algorithm (such as FLANN or BFMatcher), the identified feature points are matched to ensure consistent recognition of the same features in different frames. The matching results are screened to retain key points with high matching degrees and remove noise and false matches. The matching degree threshold is set to 0.75. If the matching score of a certain feature point in different frames is lower than this threshold, it is removed. Finally, 5 to 8 stable key reference points are retained. The coordinate information of multiple feature points identified in the mining area is recorded in the database, including the image coordinates and corresponding actual space coordinates of each feature point for subsequent analysis and modeling. A suitable 3D reconstruction algorithm (such as structured light or stereo vision) is selected to perform 3D structure analysis on the identified feature points. This algorithm should be able to deduce the 3D structure based on 2D image data. Using the stereo vision method, the depth information of the feature points is calculated through multiple images, and combined with known camera parameters, a 3D point cloud is generated. The generated 3D point cloud is processed, including downsampling, noise removal, and point cloud registration. Ensure the accuracy and usability of the point cloud data, thus providing high-quality basic data for subsequent 3D modeling. The Voxel Grid filter is used to downsample the point cloud, reducing the number of points in the point cloud to 50% of the original, and the Iterative Closest Point (ICP) algorithm is used for point cloud registration. The 3D structural features of the mining area are generated through the point cloud data, usually including the overall shape of the mining area, the relative positions of the feature points, and the main structures within the mining area. The topological relationships between the feature points are defined, including adjacency relationships, connection relationships, etc. This step ensures that the relative positions and relationships of each feature point in 3D space can be clearly described. If there is a direct connection between feature point A and feature point B, it is recorded as "feature point A - feature point B", and the weight of the connection is set according to the spatial distance. A topological analysis algorithm (such as Delaunay triangulation or Voronoi diagram) is applied to perform topological analysis on the feature points to generate a topological structure diagram between the feature points. Using the Delaunay triangulation method, the feature points are connected into a triangular network to form a topological relationship diagram, ensuring that each feature point can reflect its position in space through the connection relationship. Based on the spatial topological relationship of the feature points, a multi-level 3D modeling framework is designed. This framework should be able to combine the feature points and topological relationships to generate a 3D model with a clear hierarchy. The hierarchical relationship is set as "base layer (ground), first layer (equipment), second layer (ore pile)", ensuring that the model structure is clearly stratified.Generate a multi-level three-dimensional structure model based on the topological relationship and characteristic points. Use 3D modeling software (such as Blender or SketchUp) to combine the characteristic points with the topological relationship to form a complete three-dimensional model. Transform the characteristic points and their connection relationships into three-dimensional objects, and set the corresponding colors and materials to better reflect the actual situation of the mining area. Verify the generated three-dimensional structure model to ensure that it can accurately reflect the spatial layout and characteristics of the actual mining area. Evaluate the accuracy and reliability of the model by comparing it with the field survey data. If the deviation between the characteristic points in the model and the field measurement data is less than 5%, it is recorded that the model verification is passed, and corresponding optimization adjustments are made.

[0031] In this embodiment, refer to Figure 4 , which is a schematic diagram of the detailed implementation steps of step S3. In this embodiment, the detailed implementation steps of step S3 include: Step S31: Calculate the spatial interval of the workers' operation areas according to the position bounding boxes of multiple workers; Step S32: Perform dynamic distribution deconstruction on the spatial interval of the workers' operation areas to construct a dynamic operation distribution map of the workers; Step S33: Mark the real-time positions of the workers on the three-dimensional structure model of the mining area according to the dynamic operation distribution map of the workers to obtain a real-time worker distribution mining area structure model; Step S34: Perform real-time dynamic tracking on the marked bounding boxes of multiple workers to obtain the real-time movement trajectories of each worker; Step S35: Perform instant movement mapping on the real-time worker distribution mining area structure model based on the real-time movement trajectories of each worker to construct a real-time mining area twin model.

[0032] In this embodiment, obtain the position bounding box data of multiple workers from the monitoring system. These bounding boxes should include the position coordinates of each worker and the size information of the bounding boxes. The bounding boxes are usually calibrated by depth vision recognition technology. Assume that the bounding boxes of 5 workers are detected, namely A(200, 300, 50, 70), B(250, 320, 50, 70), etc., and record the coordinates and sizes of the bounding boxes of each worker. According to the coordinates of the workers' bounding boxes, calculate the spatial interval between the operation areas of each worker. The Euclidean distance formula can be used to calculate the distance between two workers to determine whether their operation areas overlap. If the center point of worker A is (225, 335) and the center point of worker B is (275, 355), the spatial interval is calculated as: Pixels, using a dynamic distribution deconstruction algorithm, analyze the working patterns of workers in the mining area based on the spatial interval data of the workers' operation areas. Clustering algorithms such as K-means or DBSCAN can be used to identify the working distribution of workers. Set the parameters of the clustering algorithm. Assuming we choose the K-means algorithm and set the number of clusters to 3 to identify the main working areas of workers in the mining area. According to the clustering results, generate a dynamic working distribution map of workers. This distribution map should intuitively display the working areas of each worker in the mining area and their dynamic changes. The clustering results of workers show that the workers are distributed in three main areas, which are identified by different colors. The generated working distribution map can clearly show the hot spots of workers' activities. Save the generated dynamic working distribution map of workers and record information such as the number of workers and activity frequency in each area for subsequent analysis and decision-making. The recording format is "area number, number of workers, activity frequency" to ensure the integrity of the information. Obtain the constructed three-dimensional structure model of the mining area and ensure that the model can be combined with the dynamic working distribution map of workers. The three-dimensional model should include information such as the terrain, equipment, and main routes of the mining area. The feature points in the mining area model include ore piles, equipment positions, and main channels to ensure that they can be mapped to the dynamic positions of workers. According to the real-time positions of workers and the dynamic working distribution map, update the three-dimensional structure model of the mining area and mark the specific positions of workers in the model. Real-time marking can be achieved by mapping the coordinates of workers to the coordinate system of the three-dimensional model. If the real-time position of worker A is (220, 340, 0) meters, then update the position mark of this worker in the three-dimensional model to ensure that it reflects his working status in the mining area in real time. Verify the updated real-time worker distribution mining area structure model to ensure that the marking of workers' positions is accurate. Evaluate the accuracy of the marking by comparing it with the real-time data of the monitoring system. If it is confirmed through the monitoring system that the error between the actual position of worker A and the marked position in the model is less than 5%, then record it as an update success and save the updated model. Conduct real-time dynamic tracking of the bounding boxes of multiple workers and record the movement trajectories of each worker in the mining area. Technologies such as Kalman filtering or optical flow method can be used for trajectory tracking. If a worker moves from position (220, 340) to (230, 350) in the monitoring video, then record this dynamic trajectory. Store the real-time movement trajectory data in the database to ensure that the activity history of each worker can be traced. The trajectory data should include information such as timestamps and position coordinates. The recording format is "worker number, timestamp, current position" for subsequent analysis and visualization. Visualize the movement trajectories of workers to vividly reflect the activity paths of workers in the mining area. This can be achieved by drawing trajectory lines on the three-dimensional model of the mining area. According to the movement trajectory of worker A, draw a line segment connecting his various positions on the model to intuitively display his working path. Based on the real-time movement trajectories of each worker, perform an instant movement mapping on the real-time worker distribution mining area structure model. Ensure that the model can reflect the dynamic changes of workers in real time.If worker A moves 10 meters during the monitoring period, update his position in the model to ensure that his activity status is reflected in real time. Combine the real-time worker distribution with the 3D structure model of the mining area to construct a real-time mining area twin model. This model should be able to reflect the distribution and activities of workers at different time periods. The generated mining area twin model can display the real-time positions of workers, their movement trajectories, and their relationships with the mining area environment. Verify the generated real-time mining area twin model to ensure that it can accurately reflect the actual situation of the mining area. Evaluate the accuracy of the model by comparing it with on-site data. If the error between the real-time model and the actual worker positions is less than 5%, mark it as a successful verification and perform corresponding optimizations.

[0033] In this embodiment, step S4 includes the following steps: Obtain multi-dimensional environmental monitoring parameters of the mining area based on multiple sensors; Calculate the change in oxygen content based on the multi-dimensional environmental monitoring parameters to obtain a real-time oxygen content curve; Detect toxic gases in the multi-dimensional environmental monitoring parameters and extract toxic gas parameters; Identify the gas types of the toxic gas parameters, and perform gas concentration fluctuation fitting to generate a change graph of multi-type toxic gas concentrations; Infer the air safety risks based on the change graph of multi-type toxic gas concentrations and the real-time oxygen content curve, and locate the dangerous gas areas in the real-time mining area twin model, marking the dangerous gas areas.

[0034] In this embodiment, a variety of sensors are reasonably arranged in the mining area, including oxygen sensors, toxic gas sensors (such as carbon monoxide, methane, hydrogen sulfide, etc.), temperature and humidity sensors, and barometric pressure sensors. Ensure that the sensors can cover the key areas of the mining area to monitor environmental changes. Suppose 10 oxygen sensors and 5 toxic gas sensors are arranged. The sensors should have high precision and fast response capabilities. The measurement range of the oxygen sensor is 0 - 30% (volume ratio), and the accuracy is ±0.1%. Configure a data acquisition system to regularly read the data of the sensors. Set the data acquisition frequency to once per second to ensure that environmental changes can be monitored in a timely manner. The system sends the data of each sensor to the central control system. The data format is "timestamp, sensor type, measured value", and ensure that an efficient communication protocol (such as LoRa or Zigbee) is used. According to the real-time collected oxygen concentration data, calculate the change in oxygen content and generate a real-time oxygen content curve. Data smoothing techniques (such as moving average) can be used to reduce the impact of instantaneous fluctuations. If the oxygen concentration data collected within 5 minutes is 20.5%, 20.4%, 20.6%, 20.3%, 20.5%, then calculate the moving average value and generate curve data points. Plot the calculated oxygen content data as a curve graph to show the change trend of oxygen concentration. Ensure that the curve graph can intuitively reflect the real-time change of oxygen concentration. If the curve graph shows that the oxygen concentration fluctuates in the range of 20.3% to 20.6% in the past 10 minutes, record this fluctuation situation and set an alarm threshold (such as below 20.2%) for subsequent risk warning. Conduct real-time detection of toxic gases based on multi-dimensional environmental monitoring parameters. Use toxic gas sensors to monitor the concentrations of gases such as carbon monoxide, methane, and hydrogen sulfide. Set the monitoring range of the carbon monoxide sensor to 0 - 1000 ppm, and the accuracy is ±5 ppm. If the real-time monitored carbon monoxide concentration is 150 ppm, record this value. Extract and store the collected toxic gas parameters to ensure that these data can be quickly accessed and analyzed. Record the types and concentrations of toxic gases for subsequent analysis. Use gas type recognition algorithms (such as machine learning-based classifiers) to analyze the real-time monitored gas parameters to identify the types of gases. Through the trained model, if the real-time monitored gas concentration matches the known characteristics, it is identified as carbon monoxide, methane, etc. Conduct concentration fluctuation fitting on the extracted toxic gas parameters. Methods such as linear regression or polynomial fitting can be used to generate concentration change graphs of multiple types of toxic gases. If the methane concentration data monitored in the past 60 minutes is 30 ppm, 35 ppm, 32 ppm, 40 ppm, then generate the change curve of methane concentration through fitting. Plot the fitting results as a concentration change graph to show the concentration fluctuations of different toxic gases. Ensure that the graph can intuitively reflect the change trend of gas concentration. The graph shows that the methane concentration has been gradually rising in the past 1 hour, and the risk of it exceeding the safety threshold needs to be concerned about.Based on the oxygen content change curve and the toxic gas concentration change graph, establish an air safety risk assessment model. This model should consider the mutual influence of oxygen concentration and toxic gas concentration to infer the air safety risk. According to the risk assessment results, locate and mark the areas where dangerous gases may exist in the mining area. If the risk score exceeds the set threshold (such as 0.7), mark this area as a dangerous area. If the oxygen concentration is monitored to be lower than 20.2% and the methane concentration exceeds 40 ppm in a certain area, then this area is marked as "high risk". Apply the positioning result of the dangerous gas area to the real-time mining area twin model to ensure that the model can reflect the air safety risk in the mining area in real time. Highlight the dangerous gas area in red in the 3D model of the mining area and generate a report for the management to make decisions.

[0035] In this embodiment, step S5 includes the following steps: Perform terrain morphology calculation on the real-time mining area twin model to generate terrain morphology parameters; Conduct terrain appearance semantic analysis based on the terrain morphology parameters to obtain terrain appearance semantic features; Infer dangerous terrain based on the terrain appearance semantic features to obtain dangerous terrain areas; Perform high-light visualization rendering on the real-time mining area twin model based on the dangerous terrain areas and dangerous gas areas to obtain a dangerous area warning map.

[0036] In this embodiment, data of the real-time mine twin model is obtained. This model usually contains the three-dimensional structure information of the mine area, including features such as terrain, equipment location, and ore piles. Ensure that the model can reflect the latest state of the mine area. The model may contain elevation data of the mine area, ensuring that the three-dimensional coordinates of each point can accurately describe the terrain changes in the mine area. According to the three-dimensional model of the mine area, terrain morphology parameters are calculated through algorithms. These parameters include slope, undulation, concavity and convexity, and elevation change, etc. Digital Elevation Model (DEM) technology can be used for calculation. If the elevation change range of a certain area in the mine is 100 meters, by calculating the slope change, the slope parameter is obtained as 15° and the undulation is 0.3 meters. Record all calculated parameters for subsequent analysis. Record the calculated terrain morphology parameters in the database to ensure that these data can be quickly accessed during subsequent analysis. The record format is "area number, slope, undulation, elevation change", etc. If the slope of a certain area is 10° and the undulation is 0.5 meters, it is recorded as "area A, slope 10°, undulation 0.5 meters". Select an appropriate semantic analysis method to analyze the terrain morphology parameters to extract the semantic features of the terrain appearance. Machine learning or deep learning models (such as convolutional neural networks) can be used for feature extraction. Use the trained model to analyze the terrain elevation data to identify different terrain features (such as ridges, valleys, flat areas). Organize the analysis results and extract the semantic features of the terrain appearance, including terrain type, terrain change pattern, and their corresponding risk features. Identify a certain area as a steep slope and mark its feature as "high-risk terrain". If the appearance feature of a certain area shows "steep hillside", it is recorded as "area B, type: steep slope, risk level: high". Based on the extracted semantic features of the terrain appearance, establish a dangerous terrain inference model. This model should be able to comprehensively consider the slope, terrain type, and their corresponding risk levels to infer possible dangerous terrain areas. Set threshold rules. If the slope is greater than 30° and the terrain type is "steep slope", it is marked as "high-risk terrain". Apply the inference model to analyze the terrain of the mine area to identify dangerous terrain areas. Through computer algorithms, automatically mark the areas with potential risks. If a slope of 35° and a type of "steep slope" are detected in a certain area, mark this area as a "dangerous terrain area". Integrate the previously identified dangerous gas areas with the dangerous terrain areas to form a comprehensive dangerous area dataset. Ensure that it can reflect both gas risks and terrain risks at the same time. If there is an intersection between the dangerous gas area and the dangerous terrain area, it is marked as a "high-risk area". Based on the comprehensive dangerous area data, perform high-light visualization rendering on the real-time mine twin model. Use three-dimensional visualization tools to ensure that the dangerous areas can be intuitively displayed to the management personnel. Use red highlighting to display all dangerous terrain and dangerous gas areas to ensure easy identification on the map. Generate the final dangerous area warning map and display it in real time in the mine monitoring system.Ensure that management personnel can obtain information on dangerous areas in a timely manner to facilitate the adoption of corresponding safety measures. Create a dynamic dashboard to update the status of dangerous areas in real time, ensuring that the warning map always reflects the latest safety conditions in the mining area.

[0037] In this embodiment, the specific steps of step S6 are as follows: Perform real-time prediction of the behavior trajectories of workers on the real-time mining area twin model to obtain the predicted behavior trajectories of each worker; Based on the warning map of dangerous areas, judge the regional intrusion of the predicted behavior trajectories of each worker, and mark the workers who potentially enter the dangerous areas; Based on the workers who potentially enter the dangerous areas, make real-time early warning decisions and construct a prediction strategy for dangerous area intrusion; Perform dynamic holographic visualization on the prediction strategy for dangerous area intrusion and the real-time mining area twin model to construct a holographic visualization model of mining area information.

[0038] In this embodiment, historical movement trajectory data of workers is extracted from the real-time mine area twin model. This data should include information such as the position coordinates, timestamps, and movement speeds of the workers to ensure that the behavior patterns of the workers can be reflected. The position data of each worker within the past 1 hour is collected. Assuming the movement trajectory of worker A is (220, 340), (225, 345), (230, 350), these data are recorded for subsequent analysis. Using machine learning or deep learning techniques, a worker behavior prediction model is constructed. Time series prediction algorithms such as LSTM (Long Short-Term Memory Network) can be used to capture the time dependence of the workers' movements. By inputting the past movement trajectory data of the workers, the model is trained to predict future behavior trajectories. Assuming the predicted result obtained from model training is that the coordinates of worker A at the next moment are (235, 355). According to the results output by the prediction model, the behavior prediction trajectories of each worker are generated. These trajectories should show the possible movement paths of the workers within a future period of time. If it is predicted that worker A will move to (235, 355), (240, 360), and (245, 365) within the next 5 minutes, then these predicted positions are recorded to form a complete behavior trajectory. The previously generated warning map of dangerous areas is obtained. This map should clearly identify the dangerous terrains and dangerous gas areas within the mine area. Ensure the timeliness and accuracy of the data. Assuming the slope of area B shown in the map is 35° and the methane concentration exceeds the standard, it is marked as a "high-risk area". Collision detection is performed between the behavior prediction trajectories of each worker and the dangerous areas to determine whether the workers enter the dangerous areas. Geometric methods can be used to detect whether the trajectory points intersect with the dangerous areas. If the predicted trajectory point (240, 360) of worker A intersects with dangerous area B, then this point is marked as a potential entry into the dangerous area. The detected situations of potential entry into the dangerous areas are recorded in the database for subsequent early warning decision-making. Ensure that the records include worker IDs, predicted positions, dangerous area information, etc. The record format is "worker ID, predicted position, dangerous area, timestamp" to ensure the integrity and traceability of the data. According to the situations of potential entry into the dangerous areas, a real-time early warning decision-making strategy is formulated. Ensure that this strategy can quickly respond to the situation of workers entering the dangerous areas and initiate corresponding safety measures. It is set that if a worker enters the dangerous area within a future period of time, an alarm is triggered and relevant management personnel are notified. A real-time monitoring system is configured to continuously monitor the relationship between the behavior trajectories of the workers and the dangerous areas. Once it is detected that a worker enters the dangerous area, the early warning strategy is immediately executed. If worker A enters the dangerous area in the predicted trajectory, an alarm is automatically sent to the mine area management center and this information is highlighted in the monitoring system. The execution results of the early warning decisions are recorded in the database for subsequent analysis and optimization. Ensure that all early warning events are well-documented. Appropriate holographic visualization techniques are selected to display the real-time information of the mine area and the dangerous areas. Augmented reality (AR) or virtual reality (VR) techniques can be used to achieve a more intuitive visualization effect.Adopt AR technology to overlay and display the 3D model of the mining area with real-time data, ensuring that managers can visually see the positions of workers and dangerous areas. Based on the real-time mining area twin model, the predicted trajectories of workers' behaviors, and the information of dangerous areas, construct a dynamic holographic visualization model. This model should be able to update in real time to display the safety status of the mining area. Different colors are used in the model to identify dangerous areas, and the behavior trajectories of workers are displayed through dynamic lines to ensure the real-time and accuracy of information.

[0039] In this embodiment, a visualization system for coal mine information is provided, which is used to execute the visualization method of coal mine information as described above, including: Worker vision calibration module, which is used to obtain the all-round mining area monitoring video; perform cache latency optimization and worker vision calibration, and extract multiple worker position bounding boxes; 3D structure module, which is used to identify key reference points in the mining area based on the all-round mining area monitoring video, perform 3D structure modeling, and construct a 3D structure model of the mining area; Twin model module, which is used to perform real-time dynamic tracking according to multiple worker position bounding boxes, and perform instant movement mapping on the 3D structure model of the mining area to construct a real-time mining area twin model; Toxic gas sensing module, which is used to obtain multi-dimensional environmental monitoring parameters of the mining area, perform fitting of toxic gas concentration fluctuations and positioning of dangerous gas areas, and mark dangerous gas areas; Dangerous area rendering module, which is used to infer dangerous terrains on the real-time mining area twin model, and then perform high-light visualization rendering based on the dangerous gas areas to obtain a dangerous area warning map; Real-time warning module, which is used to make real-time warning decisions based on the dangerous area warning map and perform dynamic holographic visualization to construct a holographic visualization model of mining area information.

[0040] The present invention ensures the monitoring of each worker by accurately identifying the worker's location, enabling quick understanding of their specific location in case of an emergency and reducing the occurrence of work-related injuries. By optimizing video processing cache and latency issues, the bounding box of the worker's location can be updated in real time, improving the system's response speed, reducing latency, and ensuring the smoothness of dynamic tracking. Through 3D modeling, the spatial structure of the mining area can be intuitively displayed, including complex underground passages, mine facilities, etc., helping management personnel understand the layout of the mining area. The 3D model of the mining area can serve as the basis for future design and planning, improving the efficiency of resource management and optimizing the layout of mining area facilities. By real-time modeling of the 3D structure of the mining area, management personnel can detect structural problems (such as the risk of tunnel collapse) and perform maintenance or repair in a timely manner. The twin model can update the worker's location in real time and reflect it in the 3D model, providing accurate real-time data for mining area management personnel to ensure that they can always understand the personnel dynamics in the mining area. Through real-time dynamic tracking, the twin model can help predict possible dangerous situations and take preventive measures or formulate emergency response plans in advance. The real-time tracking of the worker's location helps management personnel optimize work task arrangements and avoid personnel staying in high-risk areas. Through real-time gas concentration analysis, early warnings can be issued and measures can be taken to prevent disasters such as fires or explosions caused by toxic gas leaks. The fitting of the concentration fluctuations of toxic gases can help locate risky areas, providing a basis for personnel to take shelter and evacuate. Through 3D rendering, all dangerous areas can be intuitively displayed in the mining area twin model, greatly improving the mining area management personnel's ability to identify potential risks. The real-time marking of dangerous areas in the mining area contributes to the emergency management of the mining area, and the staff can quickly receive instructions and take measures for emergency response. Through the rendered dangerous area map, workers can quickly identify and avoid dangerous areas, enhancing the safety of mining area personnel. Through the fusion analysis of real-time data, the mining area can respond quickly to potential safety accidents, issue early warnings, and reduce the probability of accidents. Dynamic holographic visualization enables mining area management personnel to more clearly see dangerous areas and the overall safety status of the mining area, helping to make quick and accurate decisions. Through the holographic visualization model of the mining area, various types of dangers occurring in the mining area can be simulated, further enhancing the safety prevention and control ability of the mining area.

[0041] Therefore, from any perspective, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the application document are intended to be encompassed within the present invention.

[0042] The above are only specific embodiments of the present invention, enabling those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but rather to the broadest scope consistent with the principles and novel features invented herein.

Claims

1. A visualization method for coal mine information, characterized in that, It includes the following steps: Step S1: Obtain the all-round mining area monitoring video; And perform cache latency optimization and worker vision calibration, and extract the bounding boxes of multiple worker positions; Step S2: Identify the key reference points in the mining area based on the all-round mining area monitoring video, and perform three-dimensional structure modeling to construct a three-dimensional structure model of the mining area; Step S3: Perform real-time dynamic tracking based on the bounding boxes of multiple worker positions, and perform instant movement mapping on the three-dimensional structure model of the mining area to construct a real-time mining area twin model; Step S4: Obtain the multi-dimensional environmental monitoring parameters of the mining area, and perform fitting of the toxic gas concentration fluctuations and positioning of the hazardous gas areas to mark the hazardous gas areas; Step S5: Infer the dangerous terrains of the real-time mining area twin model, and then perform high-light visualization rendering based on the hazardous gas areas to obtain a warning map of the dangerous areas; Step S6: Perform real-time warning decision-making based on the warning map of the dangerous areas and perform dynamic holographic visualization to construct a holographic visualization model of the mining area information.

2. The visualization method of coal mine information according to claim 1, wherein The specific steps of Step S1 are as follows: Obtain the all-round mining area monitoring video based on multiple high-definition cameras in the mining operation area; Perform cache latency optimization on the all-round mining area monitoring video to construct a latency-optimized monitoring video; Perform deep vision detection processing on the latency-optimized monitoring video to mark each worker node; Perform real-time position calculation on the worker nodes to generate the real-time position coordinates of each worker; Perform image bounding box calibration on the latency-optimized monitoring video according to the real-time position coordinates of each worker, and extract the bounding boxes of multiple worker positions.

3. The visualization method of coal mine information according to claim 2, wherein The specific steps of performing cache latency optimization on the all-round mining area monitoring video to construct a latency-optimized monitoring video are as follows: Perform frame-by-frame video histogram brightness distribution recognition on the all-round mining area monitoring video to obtain the brightness distribution characteristics of each frame; Calculate the global average brightness value of the all-round mining area monitoring video; Perform brightness deviation position detection on the brightness distribution characteristics of each frame according to the global average brightness value, and mark the brightness deviation areas of each frame; Perform local brightness optimization on the brightness deviation areas of each frame to construct a brightness-optimized monitoring video; Perform abnormal noise point recognition on the brightness-optimized monitoring video to mark multiple video noise points; Calculate the pixel values of the video noise points; Perform adaptive filtering denoising according to the pixel values to obtain a filtering-optimized monitoring video; Perform decoding delay calculation on the filtering-optimized monitoring video to identify the video decoding delay parameters; Calculate the highest bit rate of the current network state; Perform dynamic bit rate control optimization based on the highest bit rate and the video decoding delay parameters to generate a dynamically optimized bit rate; Perform cache latency optimization on the filtering-optimized monitoring video based on the dynamically optimized bit rate to construct a latency-optimized monitoring video.

4. The visualization method of coal mine information according to claim 1, characterized in that The specific steps of Step S2 are as follows: Identify the key reference points in the mining area for the latency-optimized monitoring video to obtain multiple mining area feature points; Perform three-dimensional structure analysis of the multiple mining area feature points to generate three-dimensional structure features of the mining area; Perform feature point space topology analysis according to the multiple mining area feature points to obtain the feature point space topology relationship; Perform multi-level three-dimensional modeling on the three-dimensional structure features of the mining area based on the feature point space topology relationship to construct a three-dimensional structure model of the mining area.

5. The visualization method of coal mine information according to claim 1, characterized in that The specific steps of step S3 are as follows: Calculate the spatial interval of the workers' operation area based on the boundary boxes of multiple workers' positions; Perform dynamic distribution deconstruction on the spatial interval of the workers' operation area to construct a dynamic operation distribution map of the workers; Mark the real-time positions of the workers on the three-dimensional structure model of the mining area according to the dynamic operation distribution map of the workers to obtain a real-time worker distribution mining area structure model; Perform real-time dynamic tracking on the boundary boxes marked by multiple workers to obtain the real-time movement trajectories of each worker; Perform instant movement mapping on the real-time worker distribution mining area structure model based on the real-time movement trajectory of each worker to construct a real-time mining area twin model.

6. The visualization method of coal mine information according to claim 1, wherein The specific steps of step S4 are as follows: Obtain multi-dimensional environmental monitoring parameters of the mining area according to multiple sensors; Calculate the change in oxygen content according to the multi-dimensional environmental monitoring parameters to obtain a real-time oxygen content curve; Detect toxic gases for the multi-dimensional environmental monitoring parameters and extract toxic gas parameters; Identify the gas types of the toxic gas parameters and perform gas concentration fluctuation fitting to generate a change map of multi-type toxic gas concentrations; Infer the air safety risks for the change map of multi-type toxic gas concentrations and the real-time oxygen content curve, and locate the dangerous gas areas on the real-time mining area twin model to mark the dangerous gas areas.

7. The visualization method of coal mine information according to claim 1, characterized in that The specific steps of step S5 are as follows: Calculate the terrain morphology of the real-time mining area twin model to generate terrain morphology parameters; Perform semantic analysis of the terrain appearance based on the terrain morphology parameters to obtain terrain appearance semantic features; Infer dangerous terrains according to the terrain appearance semantic features to obtain dangerous terrain areas; Perform high-light visualization rendering on the real-time mining area twin model based on the dangerous terrain areas and the dangerous gas areas to obtain a warning map of dangerous areas.

8. The visualization method of coal mine information according to claim 1, characterized in that The specific steps of step S6 are as follows: Predict the real-time behavior trajectories of the workers on the real-time mining area twin model to obtain the behavior prediction trajectories of each worker; Judge the regional intrusion of the behavior prediction trajectories of each worker according to the warning map of dangerous areas and mark the workers who potentially enter dangerous areas; Make real-time early warning decisions based on the workers who potentially enter dangerous areas to construct a prediction strategy for dangerous area intrusion; Perform dynamic holographic visualization on the prediction strategy for dangerous area intrusion and the real-time mining area twin model to construct a holographic visualization model of mining area information.

9. A visualization system for coal mine information, characterized in that, Used to execute the visualization method of coal mine information as described in claim 1, including: A worker vision calibration module, used to obtain an all-round mining area monitoring video; and perform cache delay optimization and worker vision calibration to extract the boundary boxes of multiple workers' positions; A three-dimensional structure module, used to identify key reference points in the mining area based on the all-round mining area monitoring video and perform three-dimensional structure modeling to construct a three-dimensional structure model of the mining area; A twin model module, used to perform real-time dynamic tracking according to the boundary boxes of multiple workers' positions and perform instant movement mapping on the three-dimensional structure model of the mining area to construct a real-time mining area twin model; A toxic gas sensing module, used to obtain multi-dimensional environmental monitoring parameters of the mining area and perform toxic gas concentration fluctuation fitting and dangerous gas area location to mark the dangerous gas areas; A dangerous area rendering module, which is used to infer dangerous terrains for the real-time mine area twin model and then perform highlighted visualization rendering based on the dangerous gas area to obtain a warning map of the dangerous area; A real-time warning module, which is used to make real-time warning decisions based on the warning map of the dangerous area and perform dynamic holographic visualization to construct a holographic visualization model of the mine area information.

Images