Intelligent safety monitoring method and system based on multi-dimensional data fusion

Abstract

The invention discloses an intelligent safety monitoring method and system based on multi-dimensional data fusion. The method comprises the following steps: carrying out dynamic self-inspection on the state of industrial equipment, obtaining corresponding data of the equipment, and matching with a safety threshold to obtain the state of the equipment; an improved YOLOv7 model is used for processing the position and the physiological index of the safety helmet of the operator, and behavior data of the operator is obtained; the method comprises the following steps: acquiring environmental data by using a hierarchically deployed LoRaWAN sensor, and calculating an environmental risk index; fusing the state of the equipment, the behavior data of the personnel and the environment data by utilizing an Apache Flink real-time stream processing framework to obtain an early warning instruction; and the early warning instruction is transmitted to the self-adaptive locking control system and the personnel management platform, early warning processing is carried out, and safety monitoring is completed. According to the invention, the safety risk identification precision and response efficiency are obviously improved, and industrial safety management is promoted to be transformed from passive response to active prevention.

Description

Technical Field

[0001] This invention relates to the field of industrial safety management technology, specifically to an intelligent safety monitoring method and system based on multi-dimensional data fusion. Background Technology

[0002] The complexity and dynamism of industrial scenarios have increased significantly, revealing serious shortcomings in traditional safety management methods in the following aspects:

[0003] High reliance on manual labor, resulting in insufficient efficiency and accuracy. Current safety inspections still rely primarily on manual patrols, such as manually recording PLC parameters and visually inspecting equipment appearance for equipment status checks. Furthermore, traditional self-inspection checklists are static and fixed, unable to dynamically adjust priorities based on equipment operating status, leading to missed critical hazards. Data silos, lacking multi-dimensional correlation analysis. Equipment status monitoring, personnel positioning, and environmental parameter collection systems operate independently, with inconsistent data formats and protocols. Personnel safety monitoring methods are limited. Existing technologies are mostly based on work badge positioning or video surveillance, lacking real-time perception of personnel physiological states (such as heart rate and blood oxygen) and behavioral patterns (such as not wearing safety helmets or throwing objects from heights). Insufficient environmental monitoring coverage. Traditional methods use single-layer fixed sensor deployments, ignoring the differences in gas concentration distribution with altitude in industrial settings. Rigid early warning mechanisms, lacking intelligent decision support. Existing systems mostly trigger early warnings based on single thresholds (such as alarming when CO concentration > 50 ppm), without considering the coupling effects of multiple factors. Moreover, the lack of closed-loop control mechanisms prevents dynamic adjustment of equipment interlocking strategies based on risk levels.

[0004] In summary, existing technologies are insufficient to meet the needs of collaborative monitoring and real-time closed-loop control of "equipment-personnel-environment" in industrial scenarios. There is an urgent need for an intelligent safety monitoring system that integrates multi-dimensional data to break through data silos, improve the accuracy of risk perception, and achieve dynamic response. Summary of the Invention

[0005] The technical problem to be solved by this invention is to provide an intelligent safety monitoring method and system based on multi-dimensional data fusion, which addresses the problems of data silos, rigid early warning mechanisms, delayed response and insufficient coverage in existing industrial safety monitoring systems. This effectively improves the accuracy and response efficiency of safety monitoring in industrial scenarios, realizes accurate risk classification and adaptive response, and provides technical support for the digital transformation of industrial safety management.

[0006] To address the above technical problems, this invention proposes an intelligent security monitoring method based on multi-dimensional data fusion, comprising:

[0007] S1. Construct an intelligent safety monitoring system, which includes an intelligent safety self-inspection system, a personnel safety monitoring system, an environmental safety monitoring system, an intelligent decision support system, an adaptive interlocking control system, a personnel management platform, and a system integration and expansion module.

[0008] S2. Utilize an intelligent safety self-inspection system to dynamically self-inspect the status of industrial equipment, acquire relevant equipment data, standardize the data using edge computing nodes, and match it with safety thresholds to obtain the equipment status.

[0009] S3. Utilize the UWB (Ultra-Wideband) positioning system in the personnel safety monitoring system to locate the operator's safety helmet, and use the smart bracelet to collect the operator's physiological indicators. Then, use the improved YOLOv7 model to process the collected data and obtain the personnel's behavioral data.

[0010] S4. Collect environmental data using the layered LoRaWAN sensors deployed in the environmental safety monitoring system and calculate the environmental risk index.

[0011] S5. Transmit the results obtained in steps S2-S4 to the intelligent decision support system, and use the Apache Flink real-time stream processing framework to fuse the status of the equipment, the behavior data of the personnel, and the environmental data to obtain early warning instructions.

[0012] S6. Transmit the early warning command to the adaptive interlocking control system and personnel management platform, implement early warning processing, and complete safety monitoring.

[0013] In the S7 System Integration and Extension module, the Spring Cloud microservice architecture is used to divide services, and inter-module communication is achieved through RabbitMQ message bus and Consul service registration.

[0014] Furthermore, in step S1, the intelligent safety self-inspection system, personnel safety monitoring system, environmental safety monitoring system, adaptive interlocking control system, and personnel management platform are all connected to the intelligent decision support system.

[0015] The industrial gateway connects to the intelligent safety self-test system via an RJ45 interface.

[0016] Furthermore, in step S2, the state of the device obtained includes the following:

[0017] Construct a knowledge graph for the equipment, with entities including equipment type, failure mode and maintenance operation, and relationships including component dependency and failure association.

[0018] The priority P of the i-th device i The calculation formula is:

[0019]

[0020] Among them, C i Let F represent the historical failure frequency of the i-th device, T represent the time decay factor, and F represent the historical failure frequency of the i-th device. i S represents the weight of the i-th device, N represents the total number of devices, and S represents the weight of the ith device. t This represents the risk coefficient at the current production stage, where α, β, and γ all represent dynamic adjustment coefficients.

[0021] The equipment's relevant data includes PLC control signals, SCADA alarm information, and sensor network parameters.

[0022] Based on the Drools rule engine, real-time data is matched with a safety threshold to obtain a comprehensive deviation V, which is then validated using a random forest model. The specific expression is as follows:

[0023]

[0024] Among them, w k D represents the weight factor of the k-th data point, n represents the total number of data points, and D k S represents the measured value of the k-th data point. k λ represents the standard threshold for the k-th data point, λ represents the entropy adjustment coefficient, and Entropy represents the entropy.

[0025] Furthermore, in step S3, obtaining personnel behavior data includes the following:

[0026] An attention mechanism is introduced into the YOLOv7 model to obtain an improved YOLOv7 model.

[0027] The improved YOLOv7 model is used to track the operator's trajectory and obtain their behavioral data. The specific formula is as follows:

[0028]

[0029] Where A represents the metric value, (x,y) represents the coordinates of the operator's location, t represents time t, and m represents the total number of operators. j ,y j ) represents the coordinates of the j-th operator's location, t j Let Δx represent the time of the j-th operator, Δy represent the change in the horizontal axis, Δy represent the change in the vertical axis, Motion_Consistency represent the motion consistency measure, φ represent the spatial weight function, θ represent the time decay function, and δ represent the motion consistency coefficient.

[0030] Personnel behavior data includes not wearing a safety helmet and violating operating procedures.

[0031] Furthermore, in step S4, the calculation of the environmental risk index includes the following:

[0032] The LoRaWAN sensors were deployed at heights of 2m, 4m, 6m, and 8m, with a data sampling frequency of 1Hz. The LoRaWAN sensors include a temperature sensor and a VOC concentration sensor.

[0033] The formula for calculating the environmental risk index is:

[0034]

[0035] Where R represents the environmental risk index, T h G represents the temperature collected by the LoRaWAN sensor at layer h. h T represents the gas concentration collected by the LoRaWAN sensor at layer h. std The standard value representing temperature, G max τ represents the maximum gas concentration, Weather_Factor represents weather factors, τ represents the altitude attenuation coefficient, and η represents the weather influence factor.

[0036] Furthermore, in step S5, the warning instruction received includes the following:

[0037] The fused data is analyzed for spatiotemporal correlation using a dynamic time warping algorithm, and a graph neural network is used to dynamically adjust the risk threshold (Risk_Threshold) to obtain early warning instructions. The specific formula is as follows:

[0038] Risk_Threshold=μ·Historical_Risk+v·Real_Time_Deviation

[0039] Where μ represents the weighting coefficient of historical risk, Historical_Risk represents historical risk, v represents the weighting coefficient of real-time deviation, and Real_Time_Deviation represents real-time deviation.

[0040] Furthermore, in step S6, in the adaptive interlocking control system, the PLC valve is closed via the MQTT protocol; in the personnel management platform, access permissions are adjusted via a RESTful API.

[0041] When the overall deviation V > 0.5, the interlock control is triggered; when the environmental risk index R > 0.8, a Level III warning is triggered, relevant equipment is shut down and emergency ventilation is activated.

[0042] Furthermore, in step S7, the services in the system integration and expansion module include data acquisition services, analysis services, and early warning services.

[0043] RabbitMQ uses the Topic Exchange pattern, with the routing key format being module.sensor.type(environment.temperature.value). It implements health checks based on Consul (every 30-second interval) and automatically switches to a backup node in case of service failure.

[0044] The open API interfaces include RESTful API and GraphQL interface; the RESTful API supports OAuth 2.0 authentication, and the GraphQL interface supports complex queries (such as multi-condition filtering).

[0045] The interface is designed based on the Node-RED low-code platform, allowing users to drag and drop components to customize alert rules.

[0046] Furthermore, this invention also proposes an intelligent security monitoring system based on multi-dimensional data fusion, comprising:

[0047] The monitoring system construction module is used to build an intelligent safety monitoring system, which includes an intelligent safety self-inspection system, a personnel safety monitoring system, an environmental safety monitoring system, an intelligent decision support system, an adaptive interlocking control system, and a personnel management platform.

[0048] The equipment status acquisition module is used to perform dynamic self-inspection of the status of industrial equipment using an intelligent safety self-inspection system, acquire relevant data of the equipment, standardize the data using edge computing nodes, and match it with safety thresholds to obtain the status of the equipment.

[0049] The personnel behavior acquisition module is used to locate the safety helmets of operators using UWB in the personnel safety monitoring system and collect the physiological indicators of operators using smart bracelets. The collected data is then processed using an improved YOLOv7 model to obtain personnel behavior data.

[0050] The environmental risk index acquisition module is used to collect environmental data using LoRaWAN sensors deployed in a layered manner within the environmental safety monitoring system, and to calculate the environmental risk index.

[0051] The early warning command acquisition module is used to transmit equipment status, personnel behavior data, and environmental risk index to the intelligent decision support system. It uses the Apache Flink real-time stream processing framework to fuse equipment status, personnel behavior data, and environmental data to obtain early warning commands. The early warning commands are then transmitted to the adaptive interlocking control system and personnel management platform for early warning processing and to complete safety monitoring.

[0052] Furthermore, the present invention also proposes an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the intelligent security monitoring method based on multi-dimensional data fusion.

[0053] Furthermore, the present invention also proposes a computer-readable storage medium storing a computer program, which is executed by a processor to perform the intelligent security monitoring method based on multi-dimensional data fusion.

[0054] Compared with the prior art, the present invention, employing the above technical solution, has the following technical effects:

[0055] This invention achieves real-time collaborative perception and dynamic closed-loop control of safety risks in industrial scenarios through a multi-dimensional data fusion mechanism of equipment status, personnel behavior and environmental parameters, overcoming the blind spots in the identification of complex risks caused by data silos in traditional monitoring systems.

[0056] This invention, based on a dynamic inventory generation algorithm using knowledge graphs and a security association graph constructed using graph neural networks, overcomes the limitations of single-threshold early warning and achieves accurate risk classification and adaptive response.

[0057] This invention combines an adjustable hierarchical sensor network with an improved behavior recognition model, significantly enhancing the monitoring coverage of potential hazards such as high-altitude gas leaks and personnel violations. The modular architecture and open interface design support rapid deployment and integration with third-party systems, providing a full-chain solution for industrial safety management, from data collection and intelligent analysis to proactive prevention, effectively promoting the transformation of safety monitoring from passive response to intelligent decision-making. Attached Figure Description

[0058] Figure 1 This is a flowchart illustrating the overall implementation of the present invention. Detailed Implementation

[0059] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0060] It should be noted that, in the description of this invention patent, unless otherwise stated, "multiple" means two or more; the terms "upper", "lower", "left", "right", "inner", "outer", "front end", "rear end", "head", "tail", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing this invention patent and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation on this invention patent.

[0061] Furthermore, the terms “first,” “second,” “third,” etc., are used for descriptive purposes only and should not be interpreted as indicating or implying relative importance.

[0062] Furthermore, in the description of this invention patent, unless otherwise explicitly specified and limited, the terms "connected" and "linked" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms in this invention patent based on the specific circumstances.

[0063] To achieve the above objectives, this invention proposes an intelligent security monitoring method based on multi-dimensional data fusion, such as... Figure 1 As shown, the details are as follows:

[0064] S1. Construct an intelligent safety monitoring system, which includes an intelligent safety self-inspection system, a personnel safety monitoring system, an environmental safety monitoring system, an intelligent decision support system, an adaptive interlocking control system, a personnel management platform, and a system integration and expansion module. Specifically:

[0065] The intelligent safety self-inspection system, personnel safety monitoring system, environmental safety monitoring system, adaptive interlocking control system, and personnel management platform are all connected to the intelligent decision support system.

[0066] The industrial gateway (supporting Modbus / TCP and OPC UA protocols) connects to the intelligent safety self-test system via an RJ45 interface.

[0067] S2. Utilize an intelligent safety self-inspection system to dynamically self-inspect the status of industrial equipment, acquire relevant equipment data, and use an edge computing node (equipped with an NVIDIA Jetson Nano chip) to process this data (denoising and normalizing), then match it with safety thresholds to obtain the equipment status. Specifically:

[0068] Construct a knowledge graph for the equipment, with entities including equipment type, failure mode and maintenance operation, and relationships including component dependency and failure association.

[0069] The priority P of the i-th device i The calculation formula is:

[0070]

[0071] Among them, C i F represents the historical failure frequency of the i-th device (statistical value within the last 3 months), T represents the time decay factor (data within 3 months has a weight of 1, data within 3-6 months has a weight of 0.5), and F represents the historical failure frequency of the i-th device (statistical value within the last 3 months). i S represents the weight of the i-th device (critical devices have a weight of 2, and non-critical devices have a weight of 1), N represents the total number of devices, and S represents the weight of the ith device. t The value represents the risk coefficient of the current production stage (0.2 for the shutdown stage and 1.0 for the operation stage), and α, β, and γ all represent dynamic adjustment coefficients.

[0072] The equipment's relevant data includes PLC control signals, SCADA alarm information, and sensor network parameters.

[0073] Based on the Drools rule engine, real-time data is matched with a safety threshold to obtain a comprehensive deviation V, which is then validated using a random forest model. The specific expression is as follows:

[0074]

[0075] Among them, w k D represents the weight factor of the k-th data point, n represents the total number of data points, and D k S represents the measured value of the k-th data point. k λ represents the standard threshold for the k-th data point, λ represents the entropy adjustment coefficient (λ = 0.5), and Entropy represents the entropy.

[0076] S3. Utilize UWB (Ultra-Wideband) positioning in the personnel safety monitoring system to locate the operator's safety helmet with a positioning accuracy of ±10cm. A smart bracelet collects the operator's physiological indicators (using Bluetooth 5.0 for encrypted transmission of heart rate and blood oxygen data, with a data buffer time of <100ms). An improved YOLOv7 model is used to process the collected data to obtain the operator's behavioral data. Specifically:

[0077] An attention mechanism is introduced into the YOLOv7 model to obtain an improved YOLOv7 model.

[0078] The improved YOLOv7 model is used to track the operator's trajectory and obtain their behavioral data. The specific formula is as follows:

[0079]

[0080] Where A represents the metric value, (x,y) represents the coordinates of the operator's location, t represents time t, and m represents the total number of operators. j ,y j ) represents the coordinates of the j-th operator's location, t j Let Δx represent the time of the j-th operator, Δy represent the change in the horizontal axis, Δy represent the change in the vertical axis, Motion_Consistency represent the motion consistency measure, φ represent the spatial weight function, θ represent the time decay function, and δ represent the motion consistency coefficient.

[0081] Personnel behavior data includes not wearing a safety helmet and violating operating procedures.

[0082] S4. Collect environmental data using LoRaWAN sensors deployed in a layered manner within the environmental safety monitoring system, and calculate the environmental risk index. Specifically:

[0083] The LoRaWAN sensors were deployed at heights of 2m, 4m, 6m, and 8m, with a data sampling frequency of 1Hz. The LoRaWAN sensors include a temperature sensor and a VOC concentration sensor.

[0084] The formula for calculating the environmental risk index is:

[0085]

[0086] Where R represents the environmental risk index, T h G represents the temperature collected by the LoRaWAN sensor at layer h. h T represents the gas concentration collected by the LoRaWAN sensor at layer h. std The standard value representing temperature, G max The maximum value of the gas concentration is represented by Weather_Factor, the weather factor is represented by τ, the altitude attenuation coefficient is represented by τ = 0.2, and η is the weather influence factor (1.32 for rainy days and 1.0 for sunny days).

[0087] S5. Transmit the results obtained in steps S2-S4 to the intelligent decision support system. Using the Apache Flink real-time stream processing framework with a window size of 10 seconds, fuse device status, personnel behavior data, and environmental data to obtain early warning instructions. Specifically:

[0088] The fused data is analyzed for spatiotemporal correlation using a dynamic time warping algorithm. A three-layer graph neural network is then used to dynamically adjust the risk threshold (Risk_Threshold) to generate early warning instructions. The specific formula is as follows:

[0089] Risk_Threshold=μ·Historical_Risk+v·Real_Time_Deviation

[0090] Where μ represents the weighting coefficient of historical risk, Historical_Risk represents historical risk, v represents the weighting coefficient of real-time deviation, and Real_Time_Deviation represents real-time deviation.

[0091] μ = 0.6, v = 0.4.

[0092] S6. Transmit the early warning command to the adaptive interlocking control system and personnel management platform, implement early warning processing, and complete safety monitoring. Specifically:

[0093] In the adaptive interlocking control system, the PLC valve is closed via the MQTT protocol; in the personnel management platform, access permissions are adjusted via a RESTful API.

[0094] When the overall deviation V > 0.5, the interlock control is triggered; when the environmental risk index R > 0.8, a Level III warning is triggered, relevant equipment is shut down and emergency ventilation is activated.

[0095] In the S7 system integration and extension module, a Spring Cloud microservice architecture is used to divide services, and inter-module communication is achieved through RabbitMQ message bus and Consul service registration. Specifically:

[0096] The services include data collection, analysis, and early warning.

[0097] RabbitMQ uses the Topic Exchange pattern, with the routing key format being module.sensor.type(environment.temperature.value). It implements health checks based on Consul (every 30-second interval) and automatically switches to a backup node in case of service failure.

[0098] The open API interfaces include RESTful API and GraphQL interface; the RESTful API supports OAuth 2.0 authentication, and the GraphQL interface supports complex queries (such as multi-condition filtering).

[0099] The interface is designed based on the Node-RED low-code platform, allowing users to drag and drop components to customize alert rules.

[0100] This invention significantly improves the safety management level of industrial environments (especially coal preparation plants), realizing a shift from passive response to proactive prevention, greatly reducing the risk of safety accidents, and improving overall operational efficiency.

[0101] This invention also proposes an intelligent safety monitoring system based on multi-dimensional data fusion, including a monitoring system construction module, an equipment status acquisition module, a personnel behavior acquisition module, an environmental risk index acquisition module, an early warning command acquisition module, and a computer program that can run on a processor. It should be noted that each module in the above system corresponds to a specific step of the method provided in this invention embodiment, possessing the corresponding functional modules and beneficial effects for executing the method. Technical details not described in detail in this embodiment can be found in the method provided in this invention embodiment.

[0102] This invention also proposes an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. It should be noted that when the processor executes the computer program, it corresponds to the specific steps of the method provided in this invention, possessing the corresponding functional modules and beneficial effects for executing the method. Technical details not described in detail in this embodiment can be found in the method provided in this invention.

[0103] This invention also proposes a computer-readable storage medium storing a computer program. It should be noted that when the computer program is executed by a processor, it corresponds to the specific steps of the method provided in this invention, possessing the corresponding functional modules and beneficial effects for executing the method. Technical details not described in detail in this embodiment can be found in the method provided in this invention.

[0104] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. An intelligent security monitoring method based on multi-dimensional data fusion, characterized in that, include: S1. Construct an intelligent safety monitoring system, which includes an intelligent safety self-inspection system, a personnel safety monitoring system, an environmental safety monitoring system, an intelligent decision support system, an adaptive interlocking control system, and a personnel management platform; S2. Utilize an intelligent safety self-inspection system to dynamically self-inspect the status of industrial equipment, acquire relevant data from the equipment, standardize the data using edge computing nodes, and match it with safety thresholds to obtain the status of the equipment. S3. Use the UWB positioning of the operator's safety helmet in the personnel safety monitoring system, and the smart bracelet to collect the operator's physiological indicators. Use the improved YOLOv7 model to process the collected data and obtain the personnel's behavioral data. S4. Collect environmental data using the layered LoRaWAN sensors deployed in the environmental safety monitoring system and calculate the environmental risk index; S5. Transmit the results obtained in steps S2-S4 to the intelligent decision support system, and use the Apache Flink real-time stream processing framework to fuse the status of the equipment, the behavior data of the personnel and the environmental data to obtain early warning instructions. S6. Transmit the early warning command to the adaptive interlocking control system and personnel management platform, implement early warning processing, and complete safety monitoring.

2. The intelligent security monitoring method based on multi-dimensional data fusion according to claim 1, characterized in that, In step S1, the intelligent safety self-inspection system, personnel safety monitoring system, environmental safety monitoring system, adaptive interlocking control system, and personnel management platform are all connected to the intelligent decision support system. The industrial gateway connects to the intelligent safety self-test system via an RJ45 interface.

3. The intelligent security monitoring method based on multi-dimensional data fusion according to claim 1, characterized in that, In step S2, the status of the device is obtained, including the following: Construct a knowledge graph for the equipment, where entities include equipment type, failure mode and maintenance operation, and relationships include component dependency and failure association. The priority P of the i-th device i The calculation formula is: Among them, C i Let F represent the historical failure frequency of the i-th device, T represent the time decay factor, and F represent the historical failure frequency of the i-th device. i S represents the weight of the i-th device, N represents the total number of devices, and S represents the weight of the ith device. t This represents the risk coefficient at the current production stage, where α, β, and γ all represent dynamic adjustment coefficients. The relevant data from the equipment includes PLC control signals, SCADA alarm information, and sensor network parameters; Based on the Drools rule engine, real-time data is matched with a safety threshold to obtain a comprehensive deviation V, which is then validated using a random forest model. The specific expression is as follows: Among them, w k D represents the weight factor of the k-th data point, n represents the total number of data points, and D k S represents the measured value of the k-th data point. k λ represents the standard threshold for the k-th data point, λ represents the entropy adjustment coefficient, and Entropy represents the entropy.

4. The intelligent security monitoring method based on multi-dimensional data fusion according to claim 1, characterized in that, In step S3, the behavioral data of the personnel obtained includes the following: An attention mechanism is introduced into the YOLOv7 model to obtain an improved YOLOv7 model; The improved YOLOv7 model is used to track the operator's trajectory and obtain their behavioral data. The specific formula is as follows: Where A represents the metric value, (x,y) represents the coordinates of the operator's location, t represents time t, and m represents the total number of operators. j ,y j ) represents the coordinates of the j-th operator's location, t j Δx represents the time of the j-th operator, Δy represents the change in the horizontal axis, Δy represents the change in the vertical axis, Motion_Consistency represents the motion consistency measure, φ represents the spatial weight function, θ represents the time decay function, and δ represents the motion consistency coefficient. Personnel behavior data includes not wearing a safety helmet and violating operating procedures.

5. The intelligent security monitoring method based on multi-dimensional data fusion according to claim 3, characterized in that, In step S4, the calculation of the environmental risk index includes the following: The LoRaWAN sensors were deployed at heights of 2m, 4m, 6m, and 8m, with a data sampling frequency of 1Hz. The LoRaWAN sensors included a temperature sensor and a VOC concentration sensor. The formula for calculating the environmental risk index is: Where R represents the environmental risk index, T h G represents the temperature collected by the LoRaWAN sensor at layer h. h T represents the gas concentration collected by the LoRaWAN sensor at layer h. std The standard value representing temperature, G max τ represents the maximum gas concentration, Weather_Factor represents weather factors, τ represents the altitude attenuation coefficient, and η represents the weather influence factor.

6. The intelligent security monitoring method based on multi-dimensional data fusion according to claim 1, characterized in that, In step S5, the warning instruction received includes the following: The fused data is analyzed for spatiotemporal correlation using a dynamic time warping algorithm, and a graph neural network is used to dynamically adjust the risk threshold (Risk_Threshold) to obtain early warning instructions. The specific formula is as follows: Risk_Threshold=μ·Historical_Risk+v·Real_Time_Deviation Where μ represents the weighting coefficient of historical risk, Historical_Risk represents historical risk, v represents the weighting coefficient of real-time deviation, and Real_Time_Deviation represents real-time deviation.

7. The intelligent security monitoring method based on multi-dimensional data fusion according to claim 5, characterized in that, In step S6, when the comprehensive deviation V>0.5, the interlocking control is triggered; when the environmental risk index R>0.8, a Level III warning is triggered, the relevant equipment is shut down and emergency ventilation is started.

8. A system applied to the intelligent security monitoring method based on multi-dimensional data fusion as described in claim 1, characterized in that, include: The monitoring system construction module is used to build an intelligent safety monitoring system, which includes an intelligent safety self-inspection system, a personnel safety monitoring system, an environmental safety monitoring system, an intelligent decision support system, an adaptive interlocking control system, and a personnel management platform. The equipment status acquisition module is used to perform dynamic self-inspection of the status of industrial equipment using the intelligent safety self-inspection system, acquire the corresponding data of the equipment, standardize the data using edge computing nodes, and match it with safety thresholds to obtain the status of the equipment. The personnel behavior acquisition module is used to locate the safety helmets of operators using UWB in the personnel safety monitoring system and collect the physiological indicators of operators using smart bracelets. The collected data is then processed using an improved YOLOv7 model to obtain personnel behavior data. The environmental risk index acquisition module is used to collect environmental data using LoRaWAN sensors deployed in a layered manner in the environmental safety monitoring system and to calculate the environmental risk index. The early warning command acquisition module is used to transmit equipment status, personnel behavior data, and environmental risk index to the intelligent decision support system. It uses the Apache Flink real-time stream processing framework to fuse equipment status, personnel behavior data, and environmental data to obtain early warning commands. The early warning commands are then transmitted to the adaptive interlocking control system and personnel management platform for early warning processing and to complete safety monitoring.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the intelligent security monitoring method based on multi-dimensional data fusion as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, The computer program is executed by the processor to perform the intelligent security monitoring method based on multi-dimensional data fusion as described in any one of claims 1 to 7.

Images