Miner safety early warning system based on UWB and AI cameras

Abstract

The invention provides a miner safety early warning system based on UWB and AI cameras, and relates to the technical field of mine safety, the miner safety early warning system comprises a positioning module, a video monitoring and analysis module, a data fusion and processing module and an early warning and communication module, the positioning module is composed of a plurality of UWB base stations arranged in a mine and UWB positioning cards worn by miners, and the video monitoring and analysis module is connected with the early warning and communication module. The centimeter-level real-time positioning of miners is realized through ultra-wideband pulse signals; the UWB technology is applied, the positioning precision is improved to the centimeter level, the positions of miners can be accurately mastered, precious time is won for emergency rescue, intelligent recognition and analysis of personnel behaviors and environmental risks are achieved through the AI camera, the defects of traditional monitoring are overcome, potential safety hazards can be found in advance, and through multi-source data fusion and real-time processing, the safety of the miners is improved. The system can give an early warning at the first time when a danger occurs, thereby effectively avoiding the occurrence of an accident or reducing the harm degree of the accident.

Description

Technical Field

[0001] The present invention relates to the technical field of mine safety, and particularly to a miner safety warning system based on UWB and AI cameras. Background Art

[0002] In coal mining operations, ensuring the safety of miners is of utmost importance. However, there are many problems with traditional mine safety systems: Low positioning accuracy: The existing Bluetooth positioning technology has a large error, usually up to ±5 meters, and cannot accurately track the specific location of miners in the complex underground environment. The UWB positioning technology is single, which makes it difficult for rescue personnel to quickly and accurately find trapped miners in case of an emergency, delaying the rescue time. Isolated monitoring: The positioning system and video monitoring operate independently of each other, and there is no effective linkage between them. This results in the inability to combine personnel location information with video images to comprehensively analyze the potential risks of personnel behavior and the surrounding environment, making it difficult to detect safety hazards in advance. Lagged warning: Currently, manual inspections are mainly relied on to monitor the mine safety situation. This method cannot obtain the physiological state information of miners in real time. For example, when miners have sudden conditions such as abnormal heart rate or falling, alarms cannot be issued in time and corresponding measures cannot be taken, which may lead to serious consequences. Poor environmental adaptability: The underground environment of coal mines is harsh, with high humidity, a lot of dust, and flammable and explosive gases such as methane. Existing safety equipment is prone to failure in such an environment, with poor stability and reliability, and cannot continuously and stably provide guarantee for mine safety. In summary, the existing technology has been difficult to meet the current requirements of mine digital safety management. Therefore, the present invention proposes a miner safety warning system based on UWB and AI cameras to solve the problems existing in the prior art. Summary of the Invention

[0003] In view of the above problems, the present invention proposes a miner safety warning system based on UWB and AI cameras. This miner safety warning system based on UWB and AI cameras realizes real-time and accurate positioning of miners, intelligently analyzes personnel behavior and environmental risks, and issues warnings in a timely manner to improve the level of mine safety production.

[0004] To achieve the object of the present invention, the present invention is realized through the following technical solutions: A miner safety warning system based on UWB and AI cameras, including a positioning module, a video monitoring and analysis module, a data fusion and processing module, and a warning and communication module. The positioning module consists of multiple UWB base stations arranged in the mine and UWB positioning cards worn by miners, and realizes centimeter-level real-time positioning of miners through ultra-wideband pulse signals. The video monitoring and analysis module includes AI cameras distributed in key areas of the mine, and real-time collects and analyzes personnel behavior and environmental safety hazards. The data fusion and processing module is used to fuse and process the UWB positioning data and the video data collected by the AI cameras, and establish the correlation between the personnel position and behavior and environmental risks. The warning and communication module is used to trigger multi-level warning signals according to the fusion processing results, and realize the transmission of warning information through the underground communication network.

[0005] A further improvement lies in that: The positioning module uses UWB technology to construct a positioning system, utilizes ultra-wideband pulse signals for centimeter-level positioning, and obtains the real-time position information of miners underground.

[0006] A further improvement lies in that: When the positioning module is deployed, according to the layout and actual requirements of the mine, UWB base stations are installed at key positions in underground roadways and mining faces, and the overlapping rate of the signal coverage ranges of adjacent UWB base stations is controlled to be ≥15%.

[0007] A further improvement lies in that: The UWB positioning card integrates a low-power Bluetooth module, switches to the sleep mode when the miner is stationary, and switches to the active mode when the miner is moving, and the positioning refresh rate is ≥10Hz.

[0008] A further improvement lies in that: When the video monitoring and analysis module is deployed, the AI cameras are installed in areas with frequent personnel activities, near equipment, and areas with potential safety hazards, and the angles and parameters of the cameras are adjusted to capture the target scene.

[0009] A further improvement lies in that: The AI camera is built-in with artificial intelligence algorithms and multi-modal sensors, supports visible light and infrared imaging, and is equipped with a convolutional neural network model for intelligent analysis of the personnel behavior in the video picture, judging whether the miner has dangerous behaviors, and at the same time identifying potential safety hazards in the surrounding environment.

[0010] A further improvement lies in that: The fusion processing analysis and establishment of the correlation relationship of the data fusion and processing module include the following steps: Synchronize the UWB positioning data and the video data collected by the AI cameras through timestamps; Establish a three-dimensional coordinate system of the roadway, and map the UWB positioning coordinates to the corresponding pixel areas in the video picture; When the AI camera detects environmental risks, it generates a dynamic boundary of the risk area centered on the dangerous point with a radius of N meters in combination with UWB positioning data; Combined with UWB data, it determines the identities and quantities of the affected miners and activates the early warning mechanism; Based on historical data, it trains a risk prediction model and dynamically adjusts the determination threshold of the dangerous area.

[0011] A further improvement lies in that: the early warning and communication module includes the following early warning methods: triggering the vibration and LED flashing alarm of the UWB positioning card; playing a voice warning through the underground broadcast system; sending early warning information including location coordinates, risk types, and video data to the on-shore monitoring center.

[0012] A further improvement lies in that: based on the analysis data of the data fusion and processing module, when the early warning and communication module detects an abnormal situation, it issues an early warning through multiple early warning methods until the relevant personnel respond. At the same time, it uses the underground communication network for data transmission.

[0013] A further improvement lies in that: the underground communication network adopts a redundant design, including an underground industrial ring network, a wireless Mesh network, and a 4G / 5G emergency communication link.

[0014] The beneficial effects of the present invention are as follows: 1. The present invention applies UWB technology to improve the positioning accuracy to the centimeter level, enabling precise control of the miners' positions, winning precious time for emergency rescue. It uses an AI camera to achieve intelligent identification and analysis of personnel behavior and environmental risks, making up for the deficiencies of traditional monitoring, being able to detect potential safety hazards in advance. Through multi-source data fusion and real-time processing, the system can issue an early warning at the first moment of danger, effectively avoiding the occurrence of accidents or reducing the harm of accidents.

[0015] 2. The present invention provides comprehensive and accurate data support for mine safety management, helping managers timely understand the underground safety status, optimize management decisions, and improve the overall mine safety management level. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is the overall system architecture diagram of the present invention; Figure 2 It is the schematic diagram of underground equipment deployment of the present invention; Figure 3 It is the data processing flow chart of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0017] To deepen the understanding of the present invention, the following will further elaborate on the present invention in combination with embodiments. These embodiments are only used to explain the present invention and do not constitute a limitation on the protection scope of the present invention.

[0018] Embodiment 1 According to Figure 1 、 2 As shown in Figure 3, this embodiment proposes a miner safety warning system based on UWB and AI cameras, including a positioning module, a video monitoring and analysis module, a data fusion and processing module, and a warning and communication module. The positioning module consists of multiple UWB base stations arranged in the mine and UWB positioning cards worn by miners, and realizes centimeter-level real-time positioning of miners through ultra-wideband pulse signals; the video monitoring and analysis module includes AI cameras distributed in key areas of the mine, and real-time collects and analyzes personnel behaviors and environmental safety hazards; The data fusion and processing module is used to fuse and process the UWB positioning data and the video data collected by the AI camera, and establish the correlation between the personnel position and behaviors and environmental risks; the warning and communication module is used to trigger multi-level warning signals according to the fusion processing results, and realize the transmission of warning information through the underground communication network. Applying UWB technology, the positioning accuracy is improved to the centimeter level, and the miner's position can be accurately grasped, winning precious time for emergency rescue. Using AI cameras to realize the intelligent identification and analysis of personnel behaviors and environmental risks makes up for the deficiencies of traditional monitoring, can discover potential safety hazards in advance, and through multi-source data fusion and real-time processing, the system can issue a warning at the first time when danger occurs, effectively avoiding the occurrence of accidents or reducing the harm of accidents.

[0019] The positioning module uses UWB technology to build a positioning system, utilizes ultra-wideband pulse signals for centimeter-level positioning, and obtains the real-time position information of miners underground. When the positioning module is deployed, according to the layout and actual needs of the mine, UWB base stations are installed at key positions in underground roadways and working faces, and the signal coverage overlap rate of adjacent UWB base stations is controlled to be ≥ 15%. The UWB positioning card integrates a low-power Bluetooth module, switches to the sleep mode when the miner is stationary, and switches to the active mode when the miner is moving, and the positioning refresh rate is ≥ 10Hz. The TDOA / AOA fusion positioning algorithm is used to improve the positioning accuracy in complex environments and solve the pain point of multipath effects. An adaptive sampling strategy is designed in combination with behavioral characteristics to reduce the overall power consumption. Ensure the positioning continuity when the UWB signal fails and improve the system robustness. An anti-interference mechanism is designed for the extreme working conditions in the mine to ensure signal stability. A multi-level security guarantee system is constructed from hardware protection to emergency protocols.

[0020] When the video monitoring and analysis module is deployed, install the AI camera in areas with frequent personnel activities, near equipment, and areas with potential safety hazards, and adjust the angle and parameters of the camera to capture the target scene. The AI camera is built with artificial intelligence algorithms and multi-modal sensors, supports visible light and infrared imaging, and is equipped with a convolutional neural network model to intelligently analyze the behavior of people in the video footage, determine whether miners are engaged in dangerous behaviors, and simultaneously identify potential safety hazards in the surrounding environment. The AI camera adopts a multi-modal fusion perception architecture, integrating a high-sensitivity CMOS visible light sensor, a non-cooled infrared thermal imaging sensor, and a ToF depth sensor to build three-dimensional environmental perception capabilities; the built-in improved convolutional neural network is trained through transfer learning, loads a mine scene-specific dataset, and realizes real-time dual-channel analysis under the acceleration of an embedded NPU: the visible light channel is based on the HRNet high-resolution pose estimation model to identify abnormal behaviors through human skeleton key point tracking; the infrared channel combines a thermal radiation threshold segmentation algorithm to detect equipment overheating, water inrush hazards, and gas accumulation, and simultaneously realizes sub-millimeter crack detection of roof fissures through a multi-spectral fusion algorithm. All recognition results are output after spatio-temporal context verification, with a system latency ≤ 150 ms, a false alarm rate < 3%, and self-learning capabilities, continuously optimizing the robustness of the model under dust interference through a distributed model update protocol.

[0021] The fusion processing analysis and establishment of association relationships of the data fusion and processing module include the following steps: Synchronize UWB positioning data and video data collected by the AI camera through timestamps; Establish a three-dimensional coordinate system for the roadway and map the UWB positioning coordinates to the corresponding pixel areas in the video footage; When the AI camera detects environmental risks, generate a dynamic boundary for the risk area centered on the dangerous point with a radius of N meters in combination with the UWB positioning data; Judge the identities and numbers of affected miners in combination with the UWB data and activate the early warning mechanism; Train a risk prediction model based on historical data and dynamically adjust the determination threshold for the dangerous area.

[0022] First, based on the PTP precise clock protocol, the UWB positioning data and the AI camera video stream are synchronized with millisecond-level timestamps. Then, a three-dimensional digital twin coordinate system of the roadway is constructed through the SLAM algorithm. The absolute coordinates (x, y, z) of the UWB are mapped in real-time to the pixel area (u, v) of the corresponding camera image using perspective projection transformation, realizing the spatial association between the personnel position and visual behavior. When the AI camera identifies environmental risk points through the YOLO-X hazard detection model, the module calls the dynamic risk assessment engine, adaptively sets the radius N according to the risk type (N ∈ [5, 20] m, N = 5 for roof fall, N = 20 for gas leakage), and generates a polygonal risk area centered on P0 that automatically deforms with the roadway topology. Subsequently, the DBSCAN clustering algorithm is used to scan the UWB positioning data in real-time, extract the IDs (bound based on the MAC address of the positioning card), quantity, and relative positions of the miners entering the risk area, and trigger hierarchical warnings according to the risk level. At the same time, the system uses the online learning mechanism to train the XGBoost risk prediction model with historical accident data, dynamically optimize the decision threshold of the radius N, and realize the adaptive perception and prediction of the mine safety situation.

[0023] The warning and communication module includes the following warning methods: triggering the vibration and LED flashing alarm of the UWB positioning card; playing voice warnings through the underground broadcast system; sending warning information including position coordinates, risk types, and video data to the on-site monitoring center. The warning and communication module is based on the analysis data of the data fusion and processing module. When an abnormal situation is detected, it issues warnings through multiple warning methods until relevant personnel respond. At the same time, it uses the underground communication network for data transmission. The underground communication network adopts a redundant design, including an underground industrial ring network, a wireless Mesh network, and a 4G / 5G emergency communication link.

[0024] Embodiment 2 According to Figure 1 、 2 、and Figure 3, this embodiment proposes a miner safety warning system based on UWB and AI cameras, including a positioning module, a video monitoring and analysis module, a data fusion and processing module, and a warning and communication module; Positioning module: A high-precision positioning system is constructed using UWB technology. UWB base stations are reasonably arranged inside the mine, and miners wear integrated UWB positioning cards. The UWB technology uses its unique ultra-wideband pulse signal to achieve centimeter-level high-precision positioning, accurately obtain the real-time position information of miners underground, and provide an accurate position basis for subsequent safety warning and rescue work.

[0025] Video Surveillance and Analysis Module: Install AI cameras in key areas of the mine to conduct real-time video surveillance of the underground scenes. The AI cameras are built-in with advanced artificial intelligence algorithms, which can intelligently analyze the behavior of personnel in the video footage, such as judging whether miners have dangerous behaviors such as abnormal walking or falling, and at the same time identify potential safety hazards in the surrounding environment, such as signs of roof fall and abnormal operation of equipment.

[0026] Data Fusion and Processing Module: Integrate and process UWB positioning data and video data collected by AI cameras. Analyze multi-source data through algorithm models to establish the correlation between personnel location, behavior, and environmental risks. For example, when the AI camera detects a risk of roof fall in a certain area and the UWB positioning shows that there are miners in this dangerous area, the system immediately activates the warning mechanism.

[0027] Warning and Communication Module: When the system detects an abnormal situation, it issues warning signals in a timely manner through various means. The warning methods include but are not limited to the sound and light alarms of the wearable devices, notifications through the underground broadcast system, and real-time transmission of warning information to the on-shore monitoring center to ensure that relevant personnel can respond quickly. At the same time, use the underground communication network to achieve stable data transmission and ensure the real-time performance and reliability of the system.

[0028] Equipment Installation and Deployment: UWB Base Station Installation: Install UWB base stations at key locations such as underground roadways and working faces according to the layout and actual needs of the mine. Ensure that the distance between the base stations is reasonable and the coverage area has no dead spots to achieve high-precision positioning of the entire mine area.

[0029] AI Camera Installation: Install AI cameras in areas with frequent personnel activities, near equipment, and areas where potential safety hazards may exist. Adjust the angle and parameters of the cameras so that they can clearly capture the target scenes and provide high-quality image data for video analysis.

[0030] Miner Equipment Provision: Equip each miner with a device integrated with a UWB positioning tag and the function of receiving warning signals, ensuring that the device is comfortable to wear, easy to operate, and can operate stably in the complex underground environment.

[0031] System Commissioning and Optimization: Positioning System Calibration: After installation, calibrate the UWB positioning system. Adjust the positioning parameters by testing at different locations to ensure that the positioning accuracy meets the design requirements.

[0032] AI Algorithm Training and Optimization: Use a large amount of underground video data to train and optimize the algorithms of the AI cameras to improve their recognition accuracy of personnel behavior and environmental risks. Continuously update the algorithm model to adapt to different underground scenes and changing situations.

[0033] System linkage test: Conduct linkage tests among various modules of the system to ensure the accurate integration of UWB positioning data and video data, the normal triggering of the warning function, and the stable and reliable communication transmission. Adjust and optimize the problems found during the test in a timely manner.

[0034] System operation and maintenance: Real-time monitoring: After the system is put into operation, arrange for special personnel to be responsible for monitoring the system interface of the on-shaft monitoring center, pay attention to the location information, personnel behavior, and environmental conditions of miners in real time, and process warning information in a timely manner.

[0035] Equipment maintenance: Regularly maintain and inspect UWB base stations, AI cameras, and the equipment worn by miners, including cleaning the equipment, replacing damaged parts, upgrading software, etc., to ensure the normal operation of the equipment.

[0036] Data management: Effectively manage the large amount of data generated by the system, regularly back up important data, and mine potential safety laws and problems through data analysis to provide a basis for further optimizing the system and improving the safety management level.

[0037] This miner safety warning system based on UWB and AI cameras applies UWB technology to improve the positioning accuracy to the centimeter level, enabling precise determination of the miners' positions and gaining precious time for emergency rescue. It uses AI cameras to achieve intelligent identification and analysis of personnel behavior and environmental risks, making up for the deficiencies of traditional monitoring, and can detect potential safety hazards in advance. Through multi-source data fusion and real-time processing, the system can issue warnings at the first moment of danger, effectively avoiding the occurrence of accidents or reducing the harm of accidents. Moreover, in view of the harsh underground environment, this invention conducts special design and protection treatment on the system equipment to improve the stability and reliability of the equipment, ensuring that the system can operate normally in a high-humidity, dusty, and gas environment, providing comprehensive and accurate data support for mine safety management, helping managers understand the underground safety conditions in a timely manner, optimize management decisions, and improve the overall mine safety management level.

[0038] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification only illustrates the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. A miner safety warning system based on UWB and AI cameras, including a positioning module, a video monitoring and analysis module, a data fusion and processing module, and an early warning and communication module, characterized in that: The positioning module is composed of multiple UWB base stations arranged in the mine and UWB positioning cards worn by miners, which realize centimeter-level real-time positioning of miners through ultra-wideband pulse signals; the video monitoring and analysis module includes AI cameras distributed in key areas of the mine, which collect and analyze personnel behavior and environmental safety hazards in real time; The data fusion and processing module is used to fuse the UWB positioning data with the video data collected by the AI ​​camera to establish the correlation between the personnel location and behavior and environmental risks; the early warning and communication module is used to trigger multi-level early warning signals according to the fusion processing results, and realize the transmission of early warning information through the underground communication network.

2. According to claim 1, a miner safety warning system based on UWB and AI camera is characterized in that: The positioning module adopts UWB technology to build a positioning system, uses ultra-wideband pulse signals to perform centimeter-level positioning, and obtains the real-time position information of miners underground.

3. According to claim 2, a miner safety early warning system based on UWB and AI camera is characterized in that: When the positioning module is deployed, UWB base stations are installed at key locations of underground tunnels and mining faces according to the layout of the mine and actual needs, and the overlap rate of adjacent UWB base station signal coverage ranges is controlled to be ≥15%.

4. According to claim 3, a miner safety warning system based on UWB and AI camera is characterized in that: The UWB positioning card (1) integrates a three-axis acceleration sensor and enters a light sleep mode after a 30-second inactivity timeout; (2) automatically wakes up when encountering sudden vibrations ≥3G; (3) maintains a heartbeat connection with neighboring base stations via BLE 5.0, with a packet loss rate of less than 0.1%, switches to active mode when the miner moves, and has a positioning refresh rate of ≥10Hz.

5. The miner safety warning system based on UWB and AI camera according to claim 1 is characterized in that: When the video monitoring and analysis module is deployed, the AI ​​camera is installed in areas with frequent human activities, near equipment, and in areas where there are potential safety hazards, and the angle and parameters of the camera are adjusted so that it can capture the target scene.

6. The miner safety warning system based on UWB and AI camera according to claim 5 is characterized in that: The AI ​​camera has built-in artificial intelligence algorithms and multimodal sensors, supports visible light and infrared imaging, and is equipped with a convolutional neural network model to perform intelligent analysis of human behavior in the video, determine whether miners have dangerous behavior, and identify safety hazards in the surrounding environment.

7. The miner safety warning system based on UWB and AI camera according to claim 1 is characterized by: The fusion processing analysis and establishment of association relationship of the data fusion and processing module include the following steps: Synchronize UWB positioning data and video data collected by AI cameras through timestamps; Establish a three-dimensional coordinate system for the laneway and map the UWB positioning coordinates to the corresponding pixel area in the video screen; When the AI ​​camera detects an environmental risk, it combines the UWB positioning data to generate a dynamic boundary of the risk area with a radius of N meters centered on the danger point; the calculation of the risk radius N follows: N=α·√(A / π)+β·v, where: - α is the risk type coefficient (roof fall α=1.2, gas leakage α=2.5) - A is the area of ​​the hidden danger area - β is the tunnel wind speed correction coefficient - v is the current tunnel average wind speed" Combine UWB data to determine the identity and number of affected miners and activate the early warning mechanism; Train risk prediction models based on historical data and dynamically adjust dynamic risk assessment parameters.

8. The miner safety early warning system based on UWB and AI camera according to claim 1 is characterized by: The warning and communication module includes the following warning methods: triggering the vibration and LED flashing alarm of the UWB positioning card; playing voice warnings through the underground broadcasting system; and sending warning information containing location coordinates, risk type and video data to the surface monitoring center.

9. The miner safety warning system based on UWB and AI camera according to claim 8, characterized in that: The early warning and communication module is based on the analysis data of the data fusion and processing module. When an abnormal situation is detected, an early warning is issued through various early warning methods until relevant personnel respond. At the same time, the underground communication network is used for data transmission.

10. The miner safety warning system based on UWB and AI camera according to claim 9, characterized in that: The underground communication network adopts a redundant design, including an underground industrial ring network, a wireless mesh network and a 4G / 5G emergency communication link.

11. A data processing method based on the system of any one of claims 1 to 10, characterized in that include: (a) Synchronize UWB positioning data and video stream timestamps through the PTP protocol with an error of ≤10ms; (b) Construct a three-dimensional coordinate system for the tunnel based on the SLAM algorithm, and establish a perspective projection mapping relationship between UWB coordinates and video pixels; (c) When the AI ​​camera detects environmental risks, call the dynamic risk assessment engine to calculate the risk radius N value; (d) Use the DBSCAN clustering algorithm to scan UWB positioning data in real time to determine the number and location of affected miners; (e) Trigger a graded warning mechanism based on the risk level.

12. The method according to claim 11, characterized in that In step (c): - The risk diffusion trend is predicted by the XGBoost model, and the input parameters include: * gas concentration gradient ▽C * roof stress σ_max * tunnel cross-section wind speed distribution v(x,y) - The model training uses a five-year mine accident dataset, and the test set F1-score ≥ 0.92".

Images