AIoT fire risk dynamic assessment method and system based on multi-source data

Abstract

The invention relates to the technical field of fire risk monitoring, and discloses an AIoT fire risk dynamic assessment method and system based on multi-source data, and the system comprises a fire recognition module, a fire coverage simulation module, and a fire risk assessment module. And according to the comburent space entity model, the maximum value of the combustion flame length of the target comburent and mechanical three-dimensional design software, intelligent modeling of a three-dimensional entity model of the maximum coverage space of combustion of the comburent in the fire scene is carried out, and accurate prediction of the maximum coverage space of the fire is realized. The dangerous state of the fire scene is scientifically evaluated based on the minimum value of the distance between the target fire space and the environmental inflammables and the fire safety distance threshold value in combination with numerical analysis, the fire risk is dynamically predicted based on spatial characteristics, and the credibility and accuracy of fire risk evaluation are improved.

Description

Technical Field

[0001] This invention relates to the technical field of fire risk monitoring, specifically to an AIoT-based dynamic fire risk assessment method and system based on multi-source data. Background Technology

[0002] AIoT (Artificial Intelligence of Things) = AI (Artificial Intelligence) + IoT (Internet of Things); AIoT integrates AI and IoT technologies, generating and collecting massive amounts of data from different dimensions through the Internet of Things, storing it in the cloud and at the edge, and then using big data analysis and higher forms of artificial intelligence to achieve the digitization and intelligent interconnection of everything. The integration of IoT and AI technologies ultimately aims to form an intelligent ecosystem, achieving interoperability and interconnectivity between different intelligent terminal devices, different system platforms, and different application scenarios. The development of AIoT-related technical standards and testing standards, the implementation of related technologies, and the promotion and large-scale application of typical cases are also important issues that urgently need to be addressed in the current IoT and AI fields. AIoT is widely used in fire protection; however, existing AIoT fire monitoring cannot intelligently predict the fire coverage area, nor can it scientifically assess the fire hazard status based on the fire coverage area.

[0003] Chinese invention patent application CN119647962A, published on March 18, 2025, discloses a fire risk assessment and early warning system. This system categorizes potential fire risk factors into physical and environmental factors and collects relevant data; it constructs a risk assessment data learning sample; it classifies fire risk levels; it establishes a fire risk calculation and assessment model; and it determines the building risk level by comparing the risk score obtained from the established model with the classified fire risk levels. However, the above technical solution cannot accurately predict the fire coverage area, nor can it accurately analyze the fire hazard status based on the fire coverage area. Summary of the Invention

[0004] (a) Technical problems to be solved To address the issues mentioned above regarding the inability of existing AIoT fire monitoring to intelligently predict fire coverage or scientifically assess fire hazard status based on fire coverage, this research aims to achieve the goals of dynamically monitoring fire occurrence status, intelligently predicting fire spatial coverage, and scientifically assessing fire hazard status, thereby improving the intelligence and accuracy of fire monitoring.

[0005] (II) Technical Solution This invention is achieved through the following technical solution: an AIoT-based dynamic fire risk assessment method based on multi-source data, the method comprising the following steps: S1. Collect real-time image information of the monitoring site, identify the fire occurrence status of the monitoring site based on the real-time images, and obtain fire occurrence status identification information; when no fire occurs, repeat S1 until the fire occurrence status identification result is that a fire has occurred. S2. When a fire occurs, collect text information on the type of combustibles, the oxygen content in the environment, and the wind speed in the environment to construct spatial entity model information of combustibles and spatial entity model information of flammable materials in the environment; based on the information on the type of combustibles, oxygen content, and wind speed at the fire scene, analyze the maximum length of the burning flame of the combustibles at the fire scene, obtain the maximum length of the burning flame of the target combustibles, and construct the spatial entity model information of the maximum coverage of the fire. S3. Establish the surface coordinate information of the physical model of the maximum fire coverage space and the surface coordinate information of the physical model of the environmental flammable material space, and measure the minimum distance between the target fire space and the environmental flammable material; based on the minimum distance between the fire space and the environmental flammable material at the fire scene, analyze the dangerous state of the fire scene and obtain fire hazard state analysis information.

[0006] Preferably, real-time image information of the monitoring site is collected, and the fire occurrence status of the monitoring site is identified based on the real-time images to obtain fire occurrence status identification information; when no fire has occurred, S1 is repeated until the fire occurrence status identification result indicates that a fire has occurred. The operation steps are as follows: S11. Real-time status image information of the target monitoring site is acquired online through the camera lens, and a set of real-time image information of the monitoring site is obtained. ,in Indicates the number of collections Real-time image information of the monitoring site; S12. Based on the real-time image information set of the monitoring site and the standard feature image information set of the fire site, perform fire occurrence status identification processing on the target monitoring site to obtain fire occurrence status identification information; the fire occurrence status identification information includes no fire and fire. When the fire occurrence status identification information is no fire, repeat S1 until the fire occurrence status identification information is fire.

[0007] Preferably, the fire occurrence status identification process of the target monitoring site is performed based on the real-time image information set of the monitoring site and the standard feature image information set of the fire site to obtain fire occurrence status identification information; the fire occurrence status identification information includes no fire and fire. When the fire occurrence status identification information is no fire, the operation steps of repeating S1 until the fire occurrence status identification information is fire are as follows: S121. Establish a standard feature image information set for fire scenes. ,in Indicates the first Standard characteristic image information of a fire scene; the standard characteristic image information of a fire scene refers to the standard fire scene image information set for the target monitoring site under fire conditions; S122. The ORB image matching search algorithm is used to analyze the real-time image information set of the monitoring site. The monitoring site real-time image information described in the document The images are ordered according to their real-time monitoring image numbers and the standard feature image information set of the fire scene. The fire scene standard feature image information described in the document Perform image feature matching and generate fire occurrence status identification information based on the image feature matching results; when and If image feature matching is successful, it indicates that a fire has occurred at the target monitoring site, and the fire occurrence status identification information is output as "fire has occurred". when and If no matching of image features is found, it indicates that no fire has occurred at the target monitoring site. In this case, the fire occurrence status identification information is output as "no fire has occurred". At this time, S1 is repeated until the fire occurrence status identification information is "fire has occurred".

[0008] Preferably, when a fire occurs, textual information on the type of combustible material, ambient oxygen content, and ambient wind speed is collected to construct spatial entity model information of the combustible material and spatial entity model information of the flammable materials in the environment. Based on the combustible material type, oxygen content, and wind speed information at the fire scene, the maximum flame length of the combustible material at the fire scene is analyzed to obtain the maximum flame length of the target combustible material and construct the spatial entity model information of the maximum fire coverage. The operation steps are as follows: S21. When the fire occurrence status identification information indicates that a fire has occurred, the fire IoT monitoring platform collects online information on the material type characteristics of the combustibles at the fire scene, the ambient oxygen content at the fire scene, and the ambient wind speed at the fire scene, and obtains text information on the type of combustibles, the ambient oxygen content, and the ambient wind speed, respectively. The text information on the type of combustibles indicates the material composition of the combustibles at the fire scene, including materials such as cloth, wood, paper, gasoline, alcohol, and electrical appliances. The ambient oxygen content is expressed as a percentage; the ambient wind speed is expressed in meters per second. The fire IoT monitoring platform includes any one of the following: China Telecom Smart Fire Protection Platform, China Unicom Smart Fire Protection Cloud Platform, and Smart Fire Protection Cloud Platform. The three-dimensional solid models of burning materials and flammable materials at the fire scene are collected online by using a drone equipped with a three-dimensional point cloud camera. The three-dimensional solid model information of burning materials and flammable materials in the environment are obtained respectively. The three-dimensional solid model information of burning materials represents the three-dimensional solid model information of the burning materials at the fire scene. The three-dimensional solid model information of flammable materials in the environment represents the three-dimensional solid model information of the unburned materials at the fire scene that can burn under the action of flames. S22. Based on the text information of the type of combustible, the text information of the ambient oxygen content, the text information of the ambient wind speed, and the maximum value matrix of standard flame length for different fire scenarios, the maximum value of the flame length of the combustible at the fire scene is analyzed and processed to obtain the maximum value of the flame length of the target combustible. S23. Based on the spatial entity model information of the combustible material and the maximum length of the target combustible material's flame, a three-dimensional entity model of the maximum coverage space of the combustible material at the fire scene is constructed to obtain the entity model information of the maximum coverage space of the fire.

[0009] Preferably, the steps for analyzing and processing the maximum flame length of the combustible material at the fire scene based on the textual information of the combustible material type, the textual information of the ambient oxygen content, the textual information of the ambient wind speed, and the maximum flame length matrix of standard flames in different fire scenarios to obtain the maximum flame length of the target combustible material are as follows: S221. Establish a matrix of standard flame length maxima for different fire scenarios. ,in Indicates the first The maximum standard flame length values ​​for different fire scenario types are defined as follows: The fire scenario type is an index data type formed by combining data of the type of combustible material, ambient oxygen content, and ambient wind speed at the fire scene to search for the maximum flame length of the combustible material; the maximum standard flame length values ​​for different fire scenario types represent the maximum standard flame length values ​​set for different fire scenario types; the unit for the maximum standard flame length values ​​for different fire scenario types is centimeters. S222, Combine the text information on the type of combustible material, the text information on the ambient oxygen content, and the text information on the ambient wind speed with the matrix of maximum standard flame lengths for different fire scenarios. The maximum standard flame length for different fire scenarios described in the text By matching text information about combustible material type, ambient oxygen content, and ambient wind speed, the maximum standard flame length corresponding to the text information about combustible material type, ambient oxygen content, and ambient wind speed in different fire scenarios is retrieved. The maximum value of the target combustible flame length is generated through data identification; the specific steps for generating the maximum value of the target combustible flame length are as follows: S2221. Initialize algorithm parameters, update the maximum number of iterations T, and the flame length search population size N of the Black-crowned Tern. S2222, Calculate the matrix of maximum standard flame length in the different fire scenarios. The maximum standard flame length for all different fire scenarios in the search space The fitness value of the text information on the type of combustible material, the text information on the ambient oxygen content, and the text information on the ambient wind speed; S2223. Flame length search: The migration behavior of the black tern is analyzed in the standard flame length maximum matrix for different fire scenarios. The search space was used to find the maximum standard flame length values ​​for different fire scenarios that matched the text information on the type of combustible material, the text information on the ambient oxygen content, and the text information on the ambient wind speed. Migration behaviors include conflict avoidance, aggregation, and updating; S22231, Conflict Avoidance: The standard flame length maxima matrix for different fire scenarios The search space is updated to find the position of the tern that does not collide with other terns whose flame lengths are searchable. The position update formula is as follows: ,in The flame length search for the black tern in the first position After the iteration, the standard flame length maxima matrix for different fire scenarios is obtained. New location in the search space The flame length search for the black tern in the first position Before the next iteration, the standard flame length maxima matrix for different fire scenarios was used. The current position in the search space. These represent the variable coefficients for avoiding collisions. The formulas for calculating these variable coefficients are as follows: ,in It is used for adjustment Control variables; S22232, Clustering: Clustering refers to moving closer to the optimal position among adjacent flame lengths of the terns while avoiding conflicts, i.e., in the standard flame length maxima matrix of different fire scenarios. The search space was used to find the maximum standard flame length values ​​for different fire scenarios that best matched the text information of the type of combustible material, the text information of the ambient oxygen content, and the text information of the ambient wind speed. The optimal position is calculated using the following formula: ,in This represents the matrix of maximum standard flame lengths for individual black terns in different fire scenarios. The search space was used to find the maximum standard flame length values ​​for different fire scenarios that matched the text information on the type of combustible material, the text information on the ambient oxygen content, and the text information on the ambient wind speed. Different current positions To the optimal solution position The move variable, This represents the matrix of maximum standard flame lengths for individual black terns in different fire scenarios. The search space was used to find the maximum standard flame length values ​​for different fire scenarios that best matched the text information of the type of combustible material, the text information of the ambient oxygen content, and the text information of the ambient wind speed. Location, Represents a random variable; S22233, Update: Update refers to the standard flame length maximum value matrix in different fire scenarios. The search space includes the flame length search, the tern orientation search, the optimal solution position update trajectory. The updated trajectory calculation formula is as follows: ,in Represents the matrix of standard flame length maxima in different fire scenarios. The search space was used to find the maximum standard flame length values ​​for different fire scenarios that matched the text information on the type of combustible material, the text information on the ambient oxygen content, and the text information on the ambient wind speed. The trajectory; S2224. The flame length search tern engages in attack behavior. During the migration process, the flame length search tern uses the standard flame length maxima matrix in different fire scenarios. Within the search space, it adjusts its flight altitude, speed, and attack angle to attack prey using aerial hovering behavior, which is described in the standard flame length maxima matrix for different fire scenarios. The search space was used to find the maximum standard flame length values ​​for different fire scenarios that matched the text information on the type of combustible material, the text information on the ambient oxygen content, and the text information on the ambient wind speed. The formula for calculating hovering behavior in the air is as follows: , , ,in This represents the matrix of maximum standard flame lengths for the search of black terns in different fire scenarios. The x-coordinate of the position in the search space. This represents the matrix of maximum standard flame lengths for the search of black terns in different fire scenarios. The ordinate of the position in the search space; This represents the matrix of maximum standard flame lengths for the search of black terns in different fire scenarios. The vertical coordinate of the position in the search space; This represents the radius of each spiral. Represents the angle variable between [0, 2π]. and Representing angle variables respectively The sine and cosine values; S2225. Flame length search for the position of the black tern, and the flame length search for the black tern in the standard flame length maximum matrix of different fire scenarios. The position is updated; the position update formula is as follows: ,in The flame length search for the black tern in the first position After the iteration, the standard flame length maxima matrix for different fire scenarios is obtained. Update the position in the search space; S2226. Calculate the fitness value, i.e., the maximum value matrix of standard flame length in different fire scenarios. Calculate the maximum standard flame length for all the different fire scenarios in the search space. The fitness values ​​of the text information on the type of combustible material, the text information on the ambient oxygen content, and the text information on the ambient wind speed are used to search for the maximum standard flame length matrix for different fire scenarios. The maximum standard flame length of the different fire scenarios that best matches the text information of the type of combustible material, the text information of the ambient oxygen content, and the text information of the ambient wind speed in the global search space. fitness value; S2227. When the maximum number of iterations is met, output the maximum standard flame length value of the different fire scenarios that best matches the text information of the type of combustible material, the text information of the ambient oxygen content, and the text information of the ambient wind speed. The maximum length of the target combustible flame is generated through data identification. The maximum length of the target combustible flame represents the maximum length of the flame during the combustion process of the combustible at the fire scene. The unit of the maximum length of the target combustible flame is centimeters.

[0010] Preferably, the steps for constructing a three-dimensional solid model of the maximum coverage space of the combustible material at the fire scene based on the combustible material spatial entity model information and the maximum flame length of the target combustible material are as follows: S231. Import the spatial solid model information of the combustible material into the mechanical 3D design software and run it. Use the feature thickening tool in the mechanical 3D design software to perform feature thickening processing on the surface of the spatial solid model of the combustible material based on the combustible material flame length value corresponding to the maximum combustible material flame length value, to obtain the maximum coverage space solid model information of the fire. The maximum coverage space solid model information of the fire represents the three-dimensional solid model information of the maximum space range covered by the flame during the combustion process of the combustible material at the fire scene. The mechanical 3D design software includes any one of CATIA, SOLIDWORKS, and UG.

[0011] Preferably, the following steps are taken to establish the surface coordinate information of the solid model of the maximum fire coverage space and the surface coordinate information of the solid model of the environmental flammable material space, and measure the minimum distance between the target fire space and the environmental flammable material; based on the minimum distance between the fire space and the environmental flammable material at the fire scene, the hazard status of the fire scene is analyzed, and the fire hazard status analysis information is obtained: S31. Import the environmental flammable material spatial entity model information and the fire maximum coverage spatial entity model information into the GIS system. The GIS system performs uniformly constructed geographic spatial coordinates of the environmental flammable material spatial entity model surface corresponding to the environmental flammable material spatial entity model information and the fire maximum coverage spatial entity model surface corresponding to the fire maximum coverage spatial entity model information, respectively, to obtain the coordinate information set of the fire maximum coverage spatial entity model surface. Surface coordinate information set of spatial solid model of flammable materials in the environment ;in Indicates the generated first The surface coordinate information of the maximum fire-covered spatial entity model; the surface coordinate information of the maximum fire-covered spatial entity model includes the horizontal, vertical, and triangular coordinates of the surface coordinates of the maximum fire-covered spatial entity model; wherein Indicates the generated first The surface coordinate information of a spatial entity model of an environmental flammable substance includes the horizontal, vertical, and axial coordinates of the surface of the spatial entity model of the environmental flammable substance. S32. Based on the surface coordinate information set of the physical model of the maximum fire coverage space The surface coordinate information of the physical model of the maximum fire coverage space described in the document The set of surface coordinate information of the spatial solid model of the flammable material in the environment Surface coordinate information of the spatial solid model of flammable materials in the environment described in the document By combining spatial distance formulas to measure the distance between the fire space and flammable materials in the environment, a matrix of distance values ​​between the target fire space and flammable materials in the environment is obtained. ,in This represents the surface coordinate information of the physical model of the maximum space covered by the fire. Coordinate information of the surface of the spatial solid model of the flammable material in the environment The corresponding distance between the target fire space and flammable materials in the environment; the unit of the distance between the target fire space and flammable materials in the environment is meters; S33, Calculate the distance matrix between the target fire space and flammable materials in the environment. The distance values ​​between all the target fire spaces and flammable materials in the environment are as follows. By comparing the spacing values, the distance between the target fire space and the flammable materials in the environment with the smallest spacing value is searched. The minimum distance between the target fire space and the flammable material in the environment is generated by data identification, and the unit of the minimum distance between the target fire space and the flammable material in the environment is meters; S34. Based on the minimum distance between the target fire space and flammable materials in the environment and the fire safety distance threshold, perform a hazard state analysis of the fire scene to obtain fire hazard state analysis information.

[0012] Preferably, the steps for performing hazard analysis on the fire scene based on the minimum distance between the target fire space and flammable materials in the environment and the fire safety distance threshold to obtain fire hazard analysis information are as follows: S341. Establish a fire safety distance threshold, wherein the fire safety distance threshold represents the minimum safety distance critical value set based on the spatial size, spatial structure, and types of items at the target monitoring site, under the condition that the burning material will not cause other flammable materials to burn and can block the spread of the fire in the event of a fire at the target monitoring site. S342. Compare the minimum distance between the target fire space and the flammable materials in the environment with the fire safety distance threshold, and generate fire hazard status analysis information based on the distance comparison result. When the minimum distance between the target fire space and the flammable materials in the environment is less than the fire safety distance threshold, it indicates that the continued burning of the fire at the target monitoring site will cause other flammable materials to burn and promote the spread of the fire. In this case, the fire hazard status analysis information is output as dangerous. When the minimum distance between the target fire space and the surrounding flammable materials is not less than the fire safety distance threshold, it indicates that the continued burning of the fire at the target monitoring site will not cause other flammable materials to burn, and the burning material will disappear on its own after burning, thus preventing the fire from spreading. In this case, the fire hazard status analysis information is output as safe.

[0013] A dynamic fire risk assessment system based on multi-source data is used to implement the dynamic fire risk assessment method based on multi-source data. The system includes a fire identification module, a fire coverage simulation module, and a fire hazard assessment module. The fire identification module includes a real-time image acquisition unit for the monitoring site, a standard feature image storage unit for the fire site, and a fire occurrence status identification unit. The real-time image acquisition unit at the monitoring site acquires real-time image information of the monitoring site through a camera lens; the standard feature image storage unit at the fire site stores standard feature image information of the fire site; the fire occurrence status identification unit performs fire occurrence status identification processing on the target monitoring site based on the real-time image information of the monitoring site and the standard feature image information of the fire site to obtain fire occurrence status identification information. The fire coverage simulation module includes a combustible material type information acquisition unit, an ambient oxygen content acquisition unit, an ambient wind speed acquisition unit, a combustible material spatial entity model construction unit, an ambient flammable material spatial entity model construction unit, a standard flame length maximum value storage unit for different fire scenarios, a combustible material combustion flame length maximum value matching unit, and a fire maximum coverage spatial entity model generation unit. The combustible material type information acquisition unit collects textual information about combustible material types through the fire IoT monitoring platform; the ambient oxygen content acquisition unit collects textual information about ambient oxygen content through the fire IoT monitoring platform; the ambient wind speed acquisition unit collects textual information about ambient wind speed through the fire IoT monitoring platform; the combustible material spatial entity model construction unit collects spatial entity model information about combustible materials using a drone equipped with a 3D point cloud camera; the ambient flammable material spatial entity model construction unit collects spatial entity model information about ambient flammable materials using a drone equipped with a 3D point cloud camera; and the standard flame length maximum value storage unit for different fire scenarios is used to store standard flame length maximum values ​​for different fire scenarios. The maximum flame length of the target combustible is determined by the following: The maximum flame length matching unit analyzes and processes the maximum flame length of the combustible at the fire scene based on the combustible type text information, the ambient oxygen content text information, the ambient wind speed text information, and the standard flame length maximum values ​​for different fire scenarios, to obtain the maximum flame length of the target combustible; The maximum fire coverage space entity model generation unit constructs a three-dimensional entity model of the maximum fire coverage space based on the combustible space entity model information, the maximum flame length of the target combustible, and mechanical three-dimensional design software, to obtain the maximum fire coverage space entity model information. The fire hazard assessment module includes a fire maximum coverage space entity model surface coordinate generation unit, an environmental flammable material space entity model surface coordinate generation unit, a fire space and environmental flammable material space minimum value statistical unit, a fire safety distance threshold storage unit, and a fire hazard status analysis unit. The fire maximum coverage space entity model surface coordinate generation unit generates fire maximum coverage space entity model surface coordinate information based on the fire maximum coverage space entity model information and in conjunction with the GIS system; the environmental flammable material space entity model surface coordinate generation unit generates environmental flammable material space entity model surface coordinate information based on the environmental flammable material space entity model information and in conjunction with the GIS system; the fire space and environmental flammable material minimum distance statistics unit calculates the minimum distance between the target fire space and the environmental flammable material based on the fire maximum coverage space entity model surface coordinate information and the environmental flammable material space entity model surface coordinate information, combined with numerical analysis; the fire safety distance threshold storage unit stores fire safety distance thresholds; the fire hazard state analysis unit performs hazard state analysis processing of the fire scene based on the minimum distance between the target fire space and the environmental flammable material and the fire safety distance threshold, to obtain fire hazard state analysis information.

[0014] (III) Beneficial Effects This invention provides an AIoT-based dynamic fire risk assessment method and system based on multi-source data. It offers the following advantages: 1. By efficiently acquiring real-time images of the monitoring site through the camera lens, it provides real data support for the dynamic monitoring of fire occurrence status; based on the real-time image information of the monitoring site, combined with intelligent search algorithms and fire site standard feature image information customized based on the target monitoring site, it accurately identifies the fire occurrence status of the target monitoring site, realizes efficient and reliable monitoring of fire accidents at the target monitoring site, and improves the response efficiency of fire monitoring.

[0015] Second, by accurately collecting information on the type of combustibles, ambient oxygen content, and ambient wind speed through a fire IoT monitoring platform, the fire risk at the target monitoring site can be scientifically assessed based on multi-source feature data. A three-dimensional spatial entity model of combustibles and flammable materials in the environment at the fire site can be dynamically constructed using a drone equipped with a 3D point cloud camera, providing reliable data support for scientifically assessing fire risk based on spatial features. Based on multi-source feature information of the fire site, including combustible type, ambient oxygen content, and ambient wind speed, combined with artificial intelligence algorithms and standard flame length maxima for different fire scenarios based on big data storage, the maximum flame length of combustibles at the fire site can be accurately analyzed, enabling intelligent prediction of the maximum flame length based on multi-source data fusion. Finally, based on the spatial entity model of combustibles, the maximum flame length of the target combustibles, and mechanical 3D design software, a three-dimensional entity model of the maximum coverage space of combustibles at the fire site can be intelligently modeled, enabling accurate prediction of the maximum coverage space of the fire and improving the accuracy and quality of fire risk assessment.

[0016] Third, by combining a physical model of the maximum fire coverage space and a physical model of flammable materials in the environment with a GIS system and spatial coordinate numerical analysis, the minimum distance between combustibles and flammable materials at the fire scene can be scientifically and intelligently monitored, enabling precise analysis of fire risk based on spatial distance; based on the minimum distance between the target fire space and flammable materials in the environment, the fire safety distance threshold, and combined with numerical analysis, the dangerous state of the fire scene can be scientifically assessed, enabling dynamic prediction of fire risk based on spatial characteristics, and improving the credibility and accuracy of fire risk assessment. Attached Figure Description

[0017] Figure 1 A schematic diagram of the modules of the AIoT fire risk dynamic assessment system based on multi-source data provided by the present invention; Figure 2 The flowchart of the AIoT fire risk dynamic assessment method based on multi-source data provided by the present invention. Detailed Implementation

[0018] 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. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] An example of the AIoT-based dynamic fire risk assessment method and system based on multi-source data is as follows: Example

[0020] Please see Figures 1-2 A dynamic fire risk assessment method based on multi-source data using AIoT includes the following steps: S1. Collect real-time image information of the monitoring site, identify the fire occurrence status of the monitoring site based on the real-time images, and obtain fire occurrence status identification information; when no fire occurs, repeat S1 until the fire occurrence status identification result is that a fire has occurred. S2. When a fire occurs, collect text information on the type of combustibles, the oxygen content in the environment, and the wind speed in the environment to construct spatial entity model information of combustibles and spatial entity model information of flammable materials in the environment; based on the information on the type of combustibles, oxygen content, and wind speed at the fire scene, analyze the maximum length of the burning flame of the combustibles at the fire scene, obtain the maximum length of the burning flame of the target combustibles, and construct the spatial entity model information of the maximum coverage of the fire. S3. Establish the surface coordinate information of the physical model of the maximum fire coverage space and the surface coordinate information of the physical model of the environmental flammable material space, and measure the minimum distance between the target fire space and the environmental flammable material; based on the minimum distance between the fire space and the environmental flammable material at the fire scene, analyze the dangerous state of the fire scene and obtain fire hazard state analysis information.

[0021] For further details, please refer to Figures 1-2 The system collects real-time image information from the monitoring site, identifies the fire occurrence status based on the real-time images, and obtains fire occurrence status identification information. When no fire has occurred, the process of executing S1 is repeated until the fire occurrence status identification result indicates that a fire has occurred. The steps are as follows: S11. Real-time status image information of the target monitoring site is acquired online through the camera lens, and a set of real-time image information of the monitoring site is obtained. ,in Indicates the number of collections Real-time image information of each monitoring site; S12. Based on the real-time image information set of the monitoring site and the standard feature image information set of the fire site, perform fire occurrence status identification processing on the target monitoring site to obtain fire occurrence status identification information; the fire occurrence status identification information includes no fire and fire. When the fire occurrence status identification information is no fire, repeat S1 until the fire occurrence status identification information is fire.

[0022] Based on the real-time image information set of the monitoring site and the standard feature image information set of the fire scene, the fire occurrence status identification process of the target monitoring site is performed to obtain fire occurrence status identification information. The fire occurrence status identification information includes no fire and fire occurrence. When the fire occurrence status identification information is no fire, the operation steps of repeating S1 until the fire occurrence status identification information is fire occurrence are as follows: S121. Establish a standard feature image information set for fire scenes. ,in Indicates the first Standard characteristic image information of a fire scene; Standard characteristic image information of a fire scene represents standard fire scene image information set for a target monitoring site under fire conditions; S122. The ORB image matching search algorithm is used to collect real-time image information from the monitoring site. Real-time image information of the monitoring site Based on the real-time image numbering of the monitoring site and the standard characteristic image information set of the fire site Standard feature image information of fire scene Perform image feature matching and generate fire occurrence status identification information based on the image feature matching results; when and If image feature matching is successful, it indicates that a fire has occurred at the target monitoring site, and the fire status identification information is output as "fire has occurred". when and If no matching of image features is found, it indicates that no fire has occurred at the target monitoring site, and the fire occurrence status identification information is output as "no fire has occurred". At this time, S1 is executed repeatedly until the fire occurrence status identification information is "fire has occurred".

[0023] The monitoring site real-time image acquisition unit uses a camera lens to efficiently acquire real-time images of the monitoring site, providing real data support for dynamic monitoring of fire occurrence status. The fire site standard feature image storage unit and the fire occurrence status identification unit work together to accurately identify the fire occurrence status of the target monitoring site based on real-time image information of the monitoring site, combined with intelligent search algorithms and fire site standard feature image information customized based on the target monitoring site. This enables efficient and reliable monitoring of fire accidents at the target monitoring site and improves the response efficiency of fire monitoring.

[0024] For further details, please refer to Figures 1-2 When a fire occurs, textual information on the type of combustible material, ambient oxygen content, and ambient wind speed is collected to construct spatial entity models of the combustible material and the ambient flammable materials. Based on the combustible material type, oxygen content, and wind speed information at the fire scene, the maximum flame length of the combustible material at the fire scene is analyzed. The steps to obtain the maximum flame length of the target combustible material and construct the spatial entity model information of the maximum fire coverage are as follows: S21. When the fire status identification information indicates that a fire has occurred, the fire IoT monitoring platform collects online information on the material type characteristics of the combustibles at the fire scene, the ambient oxygen content at the fire scene, and the ambient wind speed at the fire scene, and obtains text information on the combustibles type, ambient oxygen content, and ambient wind speed, respectively. The combustibles type text information indicates the material composition of the combustibles at the fire scene, including materials such as cloth, wood, paper, gasoline, alcohol, and electrical appliances. The ambient oxygen content is expressed as a percentage; the ambient wind speed is expressed in meters per second. The fire IoT monitoring platform includes any one of the following: China Telecom Smart Fire Protection Platform, China Unicom Smart Fire Protection Cloud Platform, and Smart Fire Protection Cloud Platform. The three-dimensional solid models of burning materials and flammable materials at the fire scene were collected online by using a drone equipped with a three-dimensional point cloud camera. The spatial solid model information of burning materials and the spatial solid model information of flammable materials in the environment were obtained respectively. The spatial solid model information of burning materials represents the three-dimensional solid model information of the burning materials at the fire scene. The spatial solid model information of flammable materials in the environment represents the three-dimensional solid model information of the unburned materials at the fire scene that can burn under the action of flames. S22. Based on the text information of the type of combustible, the text information of the ambient oxygen content, the text information of the ambient wind speed, and the maximum value matrix of the standard flame length of different fire scenarios, the maximum value of the flame length of the combustible at the fire scene is analyzed and processed to obtain the maximum value of the flame length of the target combustible. S23. Based on the spatial entity model information of the combustible material and the maximum length of the target combustible material's flame, construct a three-dimensional entity model of the maximum coverage space of the combustible material at the fire scene to obtain the entity model information of the maximum coverage space of the fire.

[0025] Based on textual information about the type of combustible material, ambient oxygen content, and ambient wind speed, and using a matrix of maximum standard flame lengths for different fire scenarios, the following steps are taken to analyze and process the maximum flame length of the target combustible material at the fire scene to obtain its maximum flame length: S221. Establish a matrix of standard flame length maxima for different fire scenarios. ,in Indicates the first The maximum standard flame length values ​​for different fire scenario types are defined as follows: fire scenario type is an index data type formed by combining data of the type of combustible material, ambient oxygen content, and ambient wind speed at the fire scene to search for the maximum flame length of the combustible material; the maximum standard flame length values ​​for different fire scenario types are the maximum standard flame length values ​​set for different fire scenario types; the unit of the maximum standard flame length values ​​for different fire scenario types is centimeters. S222, Combine the text information on combustible material type, ambient oxygen content, and ambient wind speed with the matrix of maximum standard flame length for different fire scenarios. Maximum standard flame length in different fire scenarios By matching text information on combustible material type, ambient oxygen content, and ambient wind speed, the maximum standard flame length for different fire scenarios corresponding to the text information on combustible material type, ambient oxygen content, and ambient wind speed is retrieved. The maximum flame length of the target combustible is generated through data identification. The specific steps for generating the maximum flame length of the target combustible are as follows: S2221. Initialize algorithm parameters, update the maximum number of iterations T, and the flame length search population size N of the Black-crowned Tern. S2222, Calculate the matrix of maximum standard flame length in different fire scenarios. Maximum standard flame length for all different fire scenarios in the search space The fitness value is related to the text information of combustible type, ambient oxygen content, and ambient wind speed. S2223. Flame length search: terns migrate based on the standard flame length maxima matrix in different fire scenarios. The search space was used to find the maximum standard flame length values ​​for different fire scenarios that matched the text information on the type of combustible material, the ambient oxygen content, and the ambient wind speed. Migration behaviors include conflict avoidance, aggregation, and updating; S22231, Conflict Avoidance: Standard Flame Length Maximum Matrix in Different Fire Scenarios The search space is updated to find the position of the tern that does not collide with other terns whose flame lengths are searchable. The position update formula is as follows: ,in The flame length search for the black tern in the first position After the second iteration, the maximum value matrix of standard flame length in different fire scenarios New location in the search space The flame length search for the black tern in the first position Before the next iteration, the standard flame length maxima matrix for different fire scenarios The current position in the search space. These represent the variable coefficients for avoiding collisions. The formulas for calculating these variable coefficients are as follows: ,in It is used for adjustment Control variables; S22232, Clustering: Clustering refers to moving closer to the optimal position among adjacent flame lengths in the search for terns while avoiding conflicts, i.e., the maximum value matrix of standard flame lengths in different fire scenarios. The search space was used to find the maximum standard flame length values ​​for different fire scenarios that best matched the text information on the type of combustible material, the ambient oxygen content, and the ambient wind speed. The optimal position is calculated using the following formula: ,in This represents the matrix of maximum standard flame lengths for individual black terns in different fire scenarios. The search space was used to find the maximum standard flame length values ​​for different fire scenarios that matched the text information on the type of combustible material, the ambient oxygen content, and the ambient wind speed. Different current positions To the optimal solution position The move variable, This represents the matrix of maximum standard flame lengths for individual black terns in different fire scenarios. The search space was used to find the maximum standard flame length values ​​for different fire scenarios that best matched the text information on the type of combustible material, the ambient oxygen content, and the ambient wind speed. Location, Represents a random variable; S22233, Update: Update refers to the matrix of maximum standard flame length values ​​in different fire scenarios. The search space includes the flame length search, the tern orientation search, the optimal solution position update trajectory. The updated trajectory calculation formula is as follows: ,in Represents the matrix of maximum standard flame lengths in different fire scenarios The search space was used to find the maximum standard flame length values ​​for different fire scenarios that matched the text information on the type of combustible material, the ambient oxygen content, and the ambient wind speed. The trajectory; S2224. The flame length search tern engages in attack behavior. During the migration process, the flame length search tern uses the standard flame length maxima matrix in different fire scenarios. Within the search space, it adjusts its flight altitude, speed, and attack angle to attack prey using aerial hovering behavior, i.e., the maximum value matrix of standard flame length in different fire scenarios. The search space was used to find the maximum standard flame length values ​​for different fire scenarios that matched the text information on the type of combustible material, the ambient oxygen content, and the ambient wind speed. The formula for calculating hovering behavior in the air is as follows: , , ,in This represents the matrix of standard flame length maxima for the search of black terns in different fire scenarios. The x-coordinate of the position in the search space. This represents the matrix of standard flame length maxima for the search of black terns in different fire scenarios. The ordinate of the position in the search space; This represents the matrix of standard flame length maxima for the search of black terns in different fire scenarios. The vertical coordinate of the position in the search space; This represents the radius of each spiral. Represents the angle variable between [0, 2π]. and Representing angle variables respectively The sine and cosine values; S2225, Flame Length Search: Update Position of the Black Terrier; Flame Length Search: Maximum Standard Flame Length Matrix for Black Terriers in Different Fire Scenarios. The position is updated; the position update formula is as follows: ,in The flame length search for the black tern in the first position After the second iteration, the maximum value matrix of standard flame length in different fire scenarios Update the position in the search space; S2226. Calculate the fitness value, i.e., the matrix of maximum standard flame length in different fire scenarios. Calculate the maximum standard flame length for all different fire scenarios in the search space. The fitness values ​​of the text information on combustible material type, ambient oxygen content, and ambient wind speed are used to search for the maximum standard flame length matrix for different fire scenarios. The maximum standard flame length for different fire scenarios that best matches the text information on combustible material type, ambient oxygen content, and ambient wind speed in the global search space. fitness value; S2227. When the maximum number of iterations is met, output the maximum standard flame length for different fire scenarios that best matches the text information on the type of combustible material, the ambient oxygen content, and the ambient wind speed. The maximum length of the target combustible flame is generated through data identification. The maximum length of the target combustible flame represents the maximum length of the flame during the combustion process of the combustible at the fire scene. The unit of the maximum length of the target combustible flame is centimeters.

[0026] The steps for constructing a three-dimensional solid model of the maximum coverage space of the burning material at the fire scene based on the spatial entity model information of the burning material and the maximum length of the target burning material flame are as follows: S231. Import the spatial solid model information of the combustible material into the mechanical 3D design software and run it. Use the feature thickening tool in the mechanical 3D design software to perform feature thickening processing on the surface of the spatial solid model of the combustible material based on the maximum value of the combustible material's flame length. This will obtain the maximum coverage space solid model information of the fire. The maximum coverage space solid model information of the fire represents the three-dimensional solid model information of the maximum space range covered by the flame during the combustion process of the combustible material at the fire scene. The mechanical 3D design software includes any one of CATIA, SOLIDWORKS, and UG.

[0027] By coordinating the combustion material type information acquisition unit, the ambient oxygen content acquisition unit, and the ambient wind speed acquisition unit, the fire IoT monitoring platform accurately collects information on combustion material type, ambient oxygen content, and ambient wind speed, enabling scientific assessment of fire risk at the target monitoring site based on multi-source feature data. The combustion material spatial entity model construction unit and the ambient flammable material spatial entity model construction unit work together to dynamically construct spatial 3D entity models of combustion materials and ambient flammable materials at the fire scene using a drone equipped with a 3D point cloud camera, providing reliable data support for scientifically assessing fire risk based on spatial features. The combustion flame length maximum value matching unit, based on the combustion material... This system combines multi-source fire scene characteristics such as fire type information, ambient oxygen content, and ambient wind speed with artificial intelligence algorithms and standard flame length maxima based on big data storage for different fire scenarios to accurately analyze the maximum flame length of combustibles at the fire scene. This enables intelligent prediction of the maximum flame length of combustibles based on multi-source data fusion at the fire scene. A fire maximum coverage space entity model generation unit intelligently models the 3D entity model of the maximum coverage space of combustibles at the fire scene based on the combustible space entity model, the target combustible flame length maxima, and mechanical 3D design software. This achieves accurate prediction of the maximum fire coverage space, improving the accuracy and quality of fire risk assessment.

[0028] For further details, please refer to Figures 1-2 The steps for establishing the surface coordinates of the fire's maximum coverage space entity model and the environmental flammable material space entity model, and measuring the minimum distance between the target fire space and the environmental flammable material, are as follows: Based on the minimum distance between the fire space and the environmental flammable material at the fire scene, the hazard status of the fire scene is analyzed, and the fire hazard status analysis information is obtained. S31. Import the spatial entity model information of flammable materials in the environment and the spatial entity model information of the maximum fire coverage into the GIS system. The GIS system performs uniform construction of the geographic spatial coordinates of the surface of the spatial entity model corresponding to the spatial entity model information of flammable materials in the environment and the surface of the spatial entity model corresponding to the maximum fire coverage into the fire coverage into the GIS system, respectively, to obtain the coordinate information set of the surface of the spatial entity model of the maximum fire coverage. Surface coordinate information set of spatial solid model of flammable materials in the environment ;in Indicates the generated first The surface coordinate information of the maximum fire-covered spatial entity model; the surface coordinate information of the maximum fire-covered spatial entity model includes the horizontal, vertical, and triangular coordinates of the surface coordinates of the maximum fire-covered spatial entity model; among which Indicates the generated first The surface coordinate information of the spatial entity model of the environmental flammable material includes the horizontal, vertical, and axial coordinates of the surface of the spatial entity model of the environmental flammable material. S32, Based on the surface coordinate information set of the physical model of the maximum fire coverage space Surface coordinate information of the largest covered space solid model in the fire Surface coordinate information set of spatial solid model of flammable materials in the environment Surface coordinate information of a solid spatial model of flammable materials in a medium environment By combining spatial distance formulas to measure the distance between the fire space and flammable materials in the environment, a matrix of distance values ​​between the target fire space and flammable materials in the environment is obtained. ,in This indicates the surface coordinate information of the solid model of the maximum space covered by the fire. Coordinate information of the surface of the solid model of flammable materials in the environment The corresponding distance between the target fire space and flammable materials in the environment; the unit of the distance between the target fire space and flammable materials in the environment is meters; S33, Matrix of distance values ​​between the target fire space and flammable materials in the environment. Distance values ​​between all target fire spaces and flammable materials in the environment By comparing the spacing values, the target fire space and the spacing value between the target fire space and the flammable materials in the environment with the smallest spacing value are searched. The minimum distance between the target fire space and the flammable materials in the environment is generated through data identification, and the unit of the minimum distance between the target fire space and the flammable materials in the environment is meters; S34. Based on the minimum distance between the target fire space and flammable materials in the environment and the fire safety distance threshold, the hazard status analysis of the fire scene is performed to obtain fire hazard status analysis information.

[0029] The steps for analyzing and processing the hazard status of a fire scene based on the minimum distance between the target fire space and flammable materials in the environment, and the fire safety distance threshold, to obtain fire hazard status analysis information are as follows: S341. Establish a fire safety distance threshold. The fire safety distance threshold represents the minimum safety distance threshold set based on the size, structure, and type of items at the target monitoring site, so that the burning material will not cause other flammable materials to burn and can prevent the fire from spreading when a fire occurs at the target monitoring site. S342. Compare the minimum distance between the target fire space and the flammable materials in the environment with the fire safety distance threshold, and generate fire hazard status analysis information based on the distance comparison results. When the minimum distance between the target fire space and the flammable materials in the environment is less than the fire safety distance threshold, it indicates that the continued burning of the fire at the target monitoring site will cause other flammable materials to burn and promote the spread of the fire. In this case, the fire hazard status analysis information is output as dangerous. When the minimum distance between the target fire space and the surrounding flammable materials is not less than the fire safety distance threshold, it means that the continued burning of the fire at the target monitoring site will not cause other flammable materials to burn, and the burning material will disappear on its own after it stops burning, thus preventing the fire from spreading. In this case, the fire hazard status analysis information is output as safe.

[0030] By coordinating the surface coordinate generation units of the maximum fire coverage space entity model, the environmental flammable material space entity model, and the minimum distance statistical unit between the fire space and the environmental flammable material, and based on the maximum fire coverage space entity model, the environmental flammable material space entity model, combined with a GIS system and spatial coordinate numerical analysis, the minimum distance between combustibles and flammable materials at the fire scene is scientifically and intelligently monitored, enabling precise analysis of fire risk based on spatial distance. The fire hazard status analysis unit, based on the minimum distance between the target fire space and the environmental flammable material, the fire safety distance threshold, and combined with numerical analysis, scientifically assesses the hazard status of the fire scene, enabling dynamic prediction of fire risk based on spatial characteristics, and improving the credibility and accuracy of fire risk assessment.

[0031] Example 2: Please see Figures 1-2 A dynamic fire risk assessment system based on multi-source data is used to implement a dynamic fire risk assessment method based on multi-source data. The system includes a fire identification module, a fire coverage simulation module, and a fire hazard assessment module. The fire identification module includes a real-time image acquisition unit for the monitoring site, a standard feature image storage unit for the fire site, and a fire occurrence status identification unit; The monitoring site real-time image acquisition unit acquires real-time image information of the monitoring site through a camera lens; the fire scene standard feature image storage unit stores fire scene standard feature image information; the fire occurrence status identification unit performs fire occurrence status identification processing on the target monitoring site based on the real-time image information of the monitoring site and the fire scene standard feature image information to obtain fire occurrence status identification information. The fire coverage simulation module includes a combustible material type information acquisition unit, an ambient oxygen content acquisition unit, an ambient wind speed acquisition unit, a combustible material spatial entity model construction unit, an ambient flammable material spatial entity model construction unit, a standard flame length maximum value storage unit for different fire scenarios, a combustible material combustion flame length maximum value matching unit, and a fire maximum coverage spatial entity model generation unit. The system comprises the following components: a combustible material type information acquisition unit, which collects textual information about combustible material types through a fire IoT monitoring platform; an ambient oxygen content acquisition unit, which collects textual information about ambient oxygen content through a fire IoT monitoring platform; an ambient wind speed acquisition unit, which collects textual information about ambient wind speed through a fire IoT monitoring platform; a combustible material spatial entity model construction unit, which collects spatial entity model information about combustible materials using a drone equipped with a 3D point cloud camera; an ambient flammable material spatial entity model construction unit, which collects spatial entity model information about ambient flammable materials using a drone equipped with a 3D point cloud camera; a standard flame length maximum value storage unit for different fire scenarios, which stores the standard flame length maximum values ​​for different fire scenarios; a combustible material combustion flame length maximum value matching unit, which analyzes and processes the maximum flame length of combustible materials at the fire scene based on the combustible material type textual information, ambient oxygen content textual information, ambient wind speed textual information, and standard flame length maximum values ​​for different fire scenarios to obtain the maximum flame length maximum value of the target combustible material; and a fire maximum coverage space entity model generation unit, which constructs a 3D entity model of the maximum coverage space of combustible materials at the fire scene based on the combustible material spatial entity model information, the target combustible material combustion flame length maximum value, and mechanical 3D design software to obtain the fire maximum coverage space entity model information. The fire hazard assessment module includes a unit for generating surface coordinates of the solid model of the maximum fire coverage space, a unit for generating surface coordinates of the solid model of the flammable environment space, a unit for statistical analysis of the minimum distance between the fire space and the flammable environment, a unit for storing fire safety distance thresholds, and a unit for analyzing fire hazard status. The system comprises the following units: a fire maximum coverage space entity model surface coordinate generation unit, which generates surface coordinate information of the fire maximum coverage space entity model based on the fire maximum coverage space entity model information and in conjunction with the GIS system; an environmental flammable material space entity model surface coordinate generation unit, which generates surface coordinate information of the environmental flammable material space entity model based on the environmental flammable material space entity model information and in conjunction with the GIS system; a fire space and environmental flammable material minimum distance statistics unit, which calculates the minimum distance between the target fire space and environmental flammable materials based on the fire maximum coverage space entity model surface coordinate information and the environmental flammable material space entity model surface coordinate information, combined with numerical analysis; a fire safety distance threshold storage unit, used to store fire safety distance thresholds; and a fire hazard status analysis unit, which performs hazard status analysis processing of the fire scene based on the minimum distance between the target fire space and environmental flammable materials and the fire safety distance threshold, to obtain fire hazard status analysis information.

[0032] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A dynamic fire risk assessment method based on multi-source data using AIoT, characterized in that, The method includes the following steps: S1. Collect real-time image information of the monitoring site, identify the fire occurrence status of the monitoring site based on the real-time images, and obtain fire occurrence status identification information; when no fire occurs, repeat S1 until the fire occurrence status identification result is that a fire has occurred. S2. When a fire occurs, collect text information on the type of combustibles, the oxygen content in the environment, and the wind speed in the environment to construct spatial entity model information of combustibles and spatial entity model information of flammable materials in the environment; based on the information on the type of combustibles, oxygen content, and wind speed at the fire scene, analyze the maximum length of the burning flame of the combustibles at the fire scene, obtain the maximum length of the burning flame of the target combustibles, and construct the spatial entity model information of the maximum coverage of the fire. S3. Establish the surface coordinate information of the physical model of the maximum fire coverage space and the surface coordinate information of the physical model of the environmental flammable material space, and measure the minimum distance between the target fire space and the environmental flammable material. Based on the minimum distance between the fire space and flammable materials in the environment at the fire scene, the dangerous state of the fire scene is analyzed, and fire hazard state analysis information is obtained.

2. The AIoT fire risk dynamic assessment method based on multi-source data according to claim 1, characterized in that: The operation steps of S1 are as follows: S11. Real-time status image information of the target monitoring site is acquired online through the camera lens, and a set of real-time image information of the monitoring site is obtained. The include ;in Indicates the number of collections Real-time image information of the monitoring site; S12. Based on the real-time image information set of the monitoring site and the standard feature image information set of the fire site, perform fire occurrence status identification processing on the target monitoring site to obtain fire occurrence status identification information; the fire occurrence status identification information includes no fire and fire. When the fire occurrence status identification information is no fire, repeat S1 until the fire occurrence status identification information is fire.

3. The AIoT fire risk dynamic assessment method based on multi-source data according to claim 2, characterized in that: The operation steps of S12 are as follows: S121. Establish a standard feature image information set for fire scenes. The include ;in Indicates the first Standard characteristic image information of a fire scene; S122, The ORB image matching search algorithm is used to... The above According to the real-time image numbering of the monitoring site, and in order with the above The above Perform image feature matching and generate fire occurrence status identification information based on the image feature matching results; when and If image feature matching is successful, the fire occurrence status identification information is output as "a fire has occurred". when and If no matching of image features is found, the fire occurrence status identification information is output as "no fire has occurred"; at this time, S1 is repeated until the fire occurrence status identification information is "fire has occurred".

4. The AIoT fire risk dynamic assessment method based on multi-source data according to claim 3, characterized in that: The operation steps of S2 are as follows: S21. When the fire occurrence status identification information indicates that a fire has occurred, the fire Internet of Things monitoring platform collects online information on the material type characteristics of the combustibles at the fire scene, the ambient oxygen content at the fire scene, and the ambient wind speed at the fire scene, and obtains text information on the type of combustibles, the ambient oxygen content, and the ambient wind speed, respectively. The three-dimensional solid models of burning materials and flammable materials at the fire scene were collected online by using a drone equipped with a three-dimensional point cloud camera to obtain spatial solid model information of burning materials and spatial solid model information of flammable materials in the environment. S22. Based on the text information of the type of combustible, the text information of the ambient oxygen content, the text information of the ambient wind speed, and the maximum value matrix of standard flame length for different fire scenarios, the maximum value of the flame length of the combustible at the fire scene is analyzed and processed to obtain the maximum value of the flame length of the target combustible. S23. Based on the spatial entity model information of the combustible material and the maximum length of the target combustible material's flame, a three-dimensional entity model of the maximum coverage space of the combustible material at the fire scene is constructed to obtain the entity model information of the maximum coverage space of the fire.

5. The AIoT fire risk dynamic assessment method based on multi-source data according to claim 4, characterized in that: The operation steps of S22 are as follows: S221. Establish a matrix of standard flame length maxima for different fire scenarios. The include ;in Indicates the first Maximum standard flame length for different fire scenarios corresponding to various fire scenario types; S222, Combine the text information of the type of combustible material, the text information of the ambient oxygen content, and the text information of the ambient wind speed with the... The above Perform text matching of combustible material type, ambient oxygen content, and ambient wind speed to search for the corresponding text information of the combustible material type, ambient oxygen content, and ambient wind speed. The maximum value of the target combustible flame length is generated through data identification; the specific steps for generating the maximum value of the target combustible flame length are as follows: S2221. Initialize algorithm parameters, update the maximum number of iterations T, and the flame length search population size N of the Black-crowned Tern. S2222, Calculation in the All of the above in the search space The fitness value of the text information on the type of combustible material, the text information on the ambient oxygen content, and the text information on the ambient wind speed; S2223, Flame length search for migration behavior of the Black-crowned Tern, in the... The search space was used to find text information that matched the type of combustible material, the ambient oxygen content, and the ambient wind speed. Migration behaviors include conflict avoidance, aggregation, and updating; S22231, Conflict Avoidance: In the above Update the position of the flame length search tern that does not collide with other flame length search terns in the search space; S22232, Clustering: Clustering refers to moving closer to the optimal position among the terns searching for adjacent flame lengths while avoiding conflict, that is, in the... The search space was used to find the text information that best matched the text information of the type of combustible, the text information of the ambient oxygen content, and the text information of the ambient wind speed. Location; S22233, Update: Update refers to the process described above. The search space includes the flame length search, the tern orientation search, the optimal solution position update trajectory. ; S2224. The Flame Length Searching Terrier engages in attack behavior. During the migration process, the Flame Length Searching Terrier uses the above-mentioned... Within the search space, it adjusts its flight altitude, speed, and attack angle to attack prey using aerial hovering maneuvers. The search space was used to find text information that matched the type of combustible material, the ambient oxygen content, and the ambient wind speed. ; S2225, Flame Length Search Black Terrier Update Location, Flame Length Search Black Terrier is described in To update the location; S2226. Calculate the fitness value, that is, in the... Calculate all of the above in the search space The fitness values ​​of the text information on the type of combustible material, the text information on the ambient oxygen content, and the text information on the ambient wind speed are used to search for the... The text information that best matches the text information of the type of combustible material, the text information of the ambient oxygen content, and the text information of the ambient wind speed in the global search space. fitness value; S2227. When the maximum number of iterations is met, output the text information that best matches the text information of the type of combustible material, the text information of the ambient oxygen content, and the text information of the ambient wind speed. The maximum value of the target combustible flame length is generated through data identification. The maximum value of the target combustible flame length represents the maximum value of the flame length during the combustion process of the combustible at the fire scene.

6. The AIoT fire risk dynamic assessment method based on multi-source data according to claim 5, characterized in that: The operation steps of S23 are as follows: S231. Import the spatial entity model information of the combustible material into the mechanical 3D design software and run it. Use the feature thickening tool in the mechanical 3D design software to perform feature thickening processing on the surface of the spatial entity model of the combustible material based on the combustible material flame length value corresponding to the maximum combustible material flame length value, so as to obtain the spatial entity model information of the maximum fire coverage space.

7. The AIoT fire risk dynamic assessment method based on multi-source data according to claim 6, characterized in that: The operation steps of S3 are as follows: S31. Import the environmental flammable material spatial entity model information and the fire maximum coverage spatial entity model information into the GIS system. The GIS system performs uniformly constructed geographic spatial coordinates of the environmental flammable material spatial entity model surface corresponding to the environmental flammable material spatial entity model information and the fire maximum coverage spatial entity model surface corresponding to the fire maximum coverage spatial entity model information, respectively, to obtain the coordinate information set of the fire maximum coverage spatial entity model surface. Surface coordinate information set of spatial solid model of flammable materials in the environment The include The include ;in Indicates the generated first The surface coordinate information of the largest spatial entity model covered by the fire; among which Indicates the generated first Surface coordinate information of a spatial solid model of an environmentally flammable material; S32, based on the above The above With the The above By combining spatial distance formulas to measure the distance between the fire space and flammable materials in the environment, a matrix of distance values ​​between the target fire space and flammable materials in the environment is obtained. The include ;in Indicates the With the The corresponding distance value between the target fire space and flammable materials in the environment; S33, regarding the above All of the above Compare the spacing values ​​and search for the one with the smallest spacing value. And through data identification, the minimum distance between the target fire space and flammable materials in the environment is generated; S34. Based on the minimum distance between the target fire space and flammable materials in the environment and the fire safety distance threshold, perform a hazard state analysis of the fire scene to obtain fire hazard state analysis information.

8. The AIoT fire risk dynamic assessment method based on multi-source data according to claim 7, characterized in that: The operation steps of S34 are as follows: S341. Establish fire safety distance thresholds; S342. Compare the minimum distance between the target fire space and the flammable materials in the environment with the fire safety distance threshold, and generate fire hazard status analysis information based on the distance comparison result. When the minimum distance between the target fire space and the flammable materials in the environment is less than the fire safety distance threshold, the fire hazard status analysis information is output as dangerous. When the minimum distance between the target fire space and flammable materials in the environment is not less than the fire safety distance threshold, the fire hazard status analysis information is output as safe.

9. A dynamic fire risk assessment system based on multi-source data, used to implement the dynamic fire risk assessment method based on multi-source data according to any one of claims 1-8, characterized in that: The system includes a fire identification module, a fire coverage simulation module, and a fire hazard assessment module.

Images