Distributed fire starting early warning method and system based on internet of things and artificial intelligence

Abstract

The application discloses a distributed fire starting early warning method and system based on an internet of things and artificial intelligence, relates to the technical field of fire starting early warning, and determines the site structure design of a fire starting monitoring site to construct a virtual three-dimensional site structure; infrared thermal imaging images and monitoring images are used to identify combustible materials in the site by combining visual analysis technology, virtual combustible material blocks are constructed based on the size and shape features of the combustible materials, and are configured in the virtual three-dimensional site structure, and a high heat block is configured according to the heat distribution features; the combustible features of each combustible material block and the positional relationship with the high heat block are comprehensively considered to calculate a combustible risk, and early warning is triggered when the risk value is greater than or equal to a threshold value; the application realizes distributed fire starting risk assessment by collecting multi-source data in real time through the internet of things and accurately analyzing the spatial relationship between combustible materials and heat sources through artificial intelligence.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fire warning, in particular to a distributed fire warning method and system based on the Internet of Things and artificial intelligence. BACKGROUND

[0002] With the acceleration of industrialization and the improvement of urbanization level, fire operation is increasingly widely used in industrial production, construction and other fields. However, fire operation involves open fire or high temperature, often accompanied by high fire risk, especially in the presence of flammable materials, fire accidents may cause serious personnel casualties and property losses. In order to reduce the risk of fire, fire warning technology has been widely concerned and researched.

[0003] The traditional fire warning method mainly relies on single devices such as smoke detectors and temperature sensors, which trigger alarms by detecting changes in smoke concentration or temperature in the environment. However, this method has significant limitations. First, the traditional method has limited ability to identify flammable materials, making it difficult to accurately determine the type, location and potential risk of flammable materials on site, especially in complex site structures. The spatial distribution and shape characteristics of flammable materials often affect the occurrence of fire and are often ignored. Second, the traditional warning is based on static threshold judgment, lacking comprehensive analysis of flammable material characteristics and dynamic environmental factors (such as heat source proximity), resulting in insufficient sensitivity and accuracy of the warning. In addition, in complex environments such as industrial plants or construction sites, the site structure is diverse and the heat source is unevenly distributed. The traditional centralized warning system is difficult to meet the needs of different scenarios, and the response speed is slow, easy to be affected by single point failure, and difficult to meet the real-time and reliability requirements of modern fire operation.

[0004] In recent years, with the development of Internet of Things and artificial intelligence technology, fire warning based on multi-source data fusion has gradually emerged. Internet of Things technology can collect environmental data such as temperature and image in real time through distributed sensor networks; while artificial intelligence technology improves the assessment ability of fire risk through image recognition and data analysis. However, there are still the following deficiencies in the existing technology: first, there is a lack of dynamic modeling method for the spatial relationship between flammable materials and heat sources, resulting in low accuracy of risk assessment; second, the intelligence level of distributed warning system is limited, making it difficult to achieve accurate warning decision in complex environment. SUMMARY

[0005] The purpose of the present application is to provide a method and system capable of accurately warning fire.

[0006] The present application discloses a distributed fire warning method based on the Internet of Things and artificial intelligence, comprising:

[0007] Step S100, determine the site structure design of the fire monitoring site, and construct a virtual three-dimensional site structure based on the site structure design;

[0008] Step S200, acquire an infrared thermal imaging image and a monitoring image of the site, perform visual analysis on the monitoring image to determine combustible materials in the site, construct a virtual combustible material block based on the size feature and shape feature of the combustible materials in the site, and configure the virtual combustible material block in the virtual three-dimensional site structure, and perform high-heat block configuration on the virtual three-dimensional site structure based on the heat distribution feature of the infrared thermal imaging image.

[0009] Step S300, determine the combustible risk of each combustible material block based on the combustible feature of each combustible material block and the position feature of the relative high-heat block, and if the combustible risk is greater than or equal to a threshold value, a pre-alarm is performed.

[0010] In some embodiments of the present application, the method of performing visual analysis on the monitoring image comprises:

[0011] Step S201, use a visual analysis algorithm to identify the combustible materials in the monitoring image, and perform edge demarcation on the identified combustible materials to form a combustible material demarcated area.

[0012] Step S202, construct a virtual combustible material block based on the shape and size of the combustible material demarcated area.

[0013] In some embodiments of the present application, the method of constructing a virtual combustible material block comprises:

[0014] Step S2021, determine a combustible material center point, construct a vertical comparison surface at the center point, determine a combustible material mapping edge and a combustible material mapping area on the vertical comparison surface, uniformly set a plurality of shape analysis points on the combustible material mapping edge, determine a center shape analysis point between the shape analysis points, calculate an analysis point relative distance of each shape analysis point relative to the center shape analysis point, construct an analysis point relative distance group in the original order, identify the analysis point relative distance group as the shape feature of the combustible material, and identify the combustible material mapping area as the size feature of the combustible material.

[0015] Step S2022, calculate the average relative distance and the relative distance variance of the analysis point relative distance in the analysis point relative distance group, and use the average relative distance, the relative distance variance, and the combustible material mapping area of the combustible material as virtual combustible material block construction conditions to construct a virtual combustible material block.

[0016] In some embodiments of the present application, the method of constructing a virtual combustible material block using the relative distance variance of the combustible material and the combustible material mapping area as virtual combustible material block construction conditions comprises:

[0017] Step S20221, grid the virtual three-dimensional scene structure, determine the plane mapping area corresponding to the over-center shape analysis point of the combustible delineation area, and mark the grid corresponding to the plane mapping area as the basic grid;

[0018] Step S20222, based on the average relative distance of the combustible, mark the several grids above and below each basic grid for expansion, so that the span of the grids in the same vertical direction falls within the preset distance ratio interval, and mark the grids marked for expansion at this time as the core grid;

[0019] Step S20223, based on the preset distance variance interval to which the relative distance variance belongs, determine the number of outer peripheral expansion layers, and mark the outer peripheral grid of the core grid for expansion based on the number of outer peripheral expansion layers, and mark the grid marked for expansion at this time as the warning grid.

[0020] In some embodiments disclosed in the present application, the method for high-heat block configuration of the virtual three-dimensional scene structure comprises:

[0021] Step S203, perform heat analysis on the infrared thermal imaging image, and delineate the continuous area in which the heat performance reaches the preset high temperature requirement as a high-heat area;

[0022] Step S204, based on the area of the high-heat area mapped in the virtual three-dimensional scene structure, construct a high-heat spherical block in the virtual three-dimensional scene structure.

[0023] In some embodiments disclosed in the present application, the method for constructing a high-heat spherical block in the virtual three-dimensional scene structure comprises:

[0024] Step S2041, determine the area center point of the high-heat area, and determine the vertical mapping area of the high-heat area relative to the vertical plane, based on the preset area interval to which the vertical mapping area belongs, determine the radius of the high-heat spherical block, and construct the high-heat spherical block with the area center point as the center.

[0025] In some embodiments disclosed in the present application, based on the combustible characteristics of each combustible block and the position characteristics relative to the high-heat block, the method for determining the combustible risk of each combustible block comprises:

[0026] Step S301, based on the type to which the combustible belongs, determine the combustible index of the combustible, and determine the respective proximity block between the combustible block and the high-heat block;

[0027] Step S302, based on the position characteristics between the respective proximity blocks of the combustible block and the high-heat block, determine the initial risk of both, and combine the combustible index to determine the combustible risk of the combustible block.

[0028] In some embodiments of the present disclosure, the method for determining the combustible risk of the combustible block further comprises:

[0029] In step S3021, a plurality of edge mapping points are randomly set on the edges of the combustible block and the high-heat block respectively, and one edge mapping point is randomly selected on the edges of the combustible block and the high-heat block respectively for combination, to obtain a plurality of edge mapping point groups.

[0030] In step S3022, the edge mapping point distance between the edge mapping points in each edge mapping point group is determined, and the minimum edge mapping point distance and the average mapping point distance are determined, and based on the average mapping point distance and the minimum edge mapping point distance, the initial risk of the combustible block and the high-heat block is determined, and the combustible index is combined to determine the combustible risk of the combustible block.

[0031] In some embodiments of the present disclosure, a distributed hot work early warning system based on Internet of Things and artificial intelligence is also disclosed, comprising:

[0032] The first module is used for determining the site structure design of the hot work monitoring site, and constructing a virtual three-dimensional site structure based on the site structure design;

[0033] The second module is used for acquiring an infrared thermal imaging image and a monitoring image of the site, performing visual analysis on the monitoring image to determine the combustible in the site, constructing a virtual combustible block based on the size feature and the shape feature of the combustible in the site, and configuring the virtual combustible block in the virtual three-dimensional site structure, and performing high-heat block configuration on the virtual three-dimensional site structure based on the heat distribution feature of the infrared thermal imaging image.

[0034] The third module is used for determining the combustible risk of each combustible block based on the combustible feature of each combustible block and the position feature relative to the high-heat block, and if the combustible risk is greater than or equal to a threshold value, a pre-alarm is performed.

[0035] The present disclosure discloses a distributed hot work early warning method and system based on Internet of Things and artificial intelligence, and relates to the technical field of hot work early warning.

[0036] The technical solutions of the present application are described in further detail below with reference to the accompanying drawings and examples. BRIEF DESCRIPTION OF DRAWINGS

[0037] Figure 1 A method step diagram of the distributed fire starting early warning method based on the Internet of Things and artificial intelligence disclosed in the embodiments of the present application. DETAILED DESCRIPTION

[0038] The technical solutions of the present application are described in further detail below with reference to the accompanying drawings and examples.

[0039] The technical solutions of the present application are described in further detail below with reference to the accompanying drawings and examples.

[0040] Embodiment:

[0041] The present application discloses a distributed fire starting early warning method based on the Internet of Things and artificial intelligence, referring to Figure 1 , comprising:

[0042] Step S100, determine the site structure design of the fire monitoring site, and construct a virtual three-dimensional site structure based on the site structure design.

[0043] The core principle of step S100 is to realize the digital collection of the physical structure of the hot work site through Internet of Things technology, and to convert this information into a virtual three-dimensional site structure using computer graphics and three-dimensional modeling techniques, providing a spatial reference framework for subsequent flammable material positioning and risk analysis. Internet of Things devices such as laser radars, cameras, and ultrasonic sensors are deployed in a distributed manner to scan the physical environment of the site in real time, generating high-precision point cloud data or two-dimensional image sequences. These data reflect the structural features of the site, such as walls, floors, ceilings, equipment, and pipelines. For example, in an industrial plant, a laser radar can scan the steel structure, mechanical equipment location, and passage layout inside the plant, generating point cloud data containing three-dimensional coordinates. Subsequently, through three-dimensional modeling algorithms such as stereoscopy, voxelization, or triangulation, the point cloud data is converted into a continuous virtual three-dimensional model. This process is similar to the construction of a Building Information Model (BIM), but with a greater emphasis on real-time and dynamic updates. The significance of constructing a virtual three-dimensional site structure lies in the fact that it is not only a static digital twin, but also can be linked with the Internet of Things sensor network to reflect spatial changes in real time. For example, if materials are temporarily stacked in the plant, sensors can detect the new obstacles and update the model to ensure that the virtual structure is consistent with reality. In addition, this step also needs to consider the functional zoning of the site (such as welding area, storage area) to provide context information for the configuration of flammable materials and heat sources in the subsequent steps. For example, suppose in a hot work site of a chemical plant, the technical personnel input the design drawings of the plant (such as CAD files), combined with the real-time data scanned by Internet of Things devices, the software automatically generates a three-dimensional model containing reaction vessels, pipelines, and storage tanks. The spatial coordinates (such as x, y, z) of each object in the model are accurately recorded, laying the foundation for the configuration of flammable material blocks and high-heat blocks in step S200. This digital spatial modeling technology solves the limitations of traditional hot work monitoring, which relies on manual inspection or two-dimensional floor plans, enabling fire risk analysis in complex environments to be carried out in three-dimensional space. Further, the construction of a virtual three-dimensional structure can also be processed through gridding (as described in step S20221), dividing the space into computable units for subsequent geometric analysis and risk assessment. The precision of the grid (such as 1 centimeter or 10 centimeters) depends on the complexity of the site and the real-time requirements of the early warning.

[0044] In step S200, infrared thermal imaging images and monitoring images of the site are obtained, visual analysis is performed on the monitoring images to determine the flammable materials in the site, virtual flammable material blocks are constructed based on the size and shape characteristics of the flammable materials in the site, and the virtual flammable material blocks are configured in the virtual three-dimensional site structure. Based on the heat distribution characteristics of the infrared thermal imaging images, high-heat blocks are configured in the virtual three-dimensional site structure.

[0045] The principle of step S200 is to achieve precise identification and spatial mapping of flammable materials and heat sources in the hot work site through multi-source data collection of the Internet of Things and image processing technology of artificial intelligence, and finally to construct flammable material blocks and high-heat blocks in the virtual three-dimensional site structure, providing data support for risk assessment. First, Internet of Things devices such as infrared thermal imagers and high-definition cameras are deployed in the site to collect infrared thermal imaging images and monitoring images respectively. Infrared thermal imaging images generate temperature distribution maps based on the thermal radiation of object surfaces, which can reflect potential high-temperature areas, such as heat diffusion around a welding point; while monitoring images provide information on the appearance of objects under visible light, such as color, texture and contour, which facilitates the identification of the type and location of flammable materials. For example, in a wood processing plant, an infrared thermal imager may detect a high-temperature area near a sawing machine, while a camera captures wooden boards and sawdust piled nearby. Then, through artificial intelligence visual analysis algorithms such as convolutional neural networks (CNN) or YOLO models, the monitoring images are processed to identify flammable materials and delineate their edges (step S201). For example, the algorithm can identify the rectangular contour of a pile of wooden boards and mark it as a flammable material delineation area. This process relies on a pre-trained model that learns the visual features of flammable materials through a large amount of labeled data (such as images of wood, paper, plastic). After identification, based on the size and shape characteristics of flammable materials (as described in steps S2021-S2022), virtual flammable material blocks are constructed. Specifically, by determining the center point and vertical mapping surface of the flammable material, calculating its mapping area and the relative distance of the edge points, and quantifying its geometric characteristics. For example, the mapping area of a wooden board may be 2 square meters, and the edge point distance analysis shows that it is a regular rectangle, and these data are converted into a virtual block. The construction of the block also considers the gridding method (steps S20221-S20223), which divides the virtual three-dimensional structure into grid cells and marks the core grid and warning grid according to the average relative distance and variance extension, ensuring that the spatial range of the block is reasonable. After completion, the virtual flammable material block is placed in the corresponding location in the virtual three-dimensional site structure, for example, the wooden board block is placed in the coordinate area next to the sawing machine. At the same time, based on the heat distribution characteristics of the infrared thermal imaging images (steps S203-S204), the high-temperature areas are analyzed and high-heat blocks are constructed. In principle, the pixel value of the infrared image is proportional to the temperature, and the high-heat area is delineated by threshold segmentation (such as continuous areas with a temperature exceeding 100°C), and then a high-heat spherical block is constructed with the center of the area as the center (step S2041). For example, if a 150°C high-temperature area is detected near the sawing machine, with a center point coordinate of (5, 3, 1), the radius is determined to be 0.5 meters according to the mapping area, then a spherical high-heat block is generated and embedded in the virtual structure. This process realizes the spatial visualization of heat sources and forms a correlation with the flammable material blocks.

[0046] Step S300, based on the flammable characteristics of each flammable group and the location characteristics of the relative high-heat block, determine the flammable risk of each flammable group, and if the flammable risk is greater than or equal to the threshold value, a pre-alarm is given.

[0047] The principle of step S300 is to analyze the inherent flammable characteristics of the flammable group and the spatial position relationship of the high-heat block, quantify the fire risk and realize intelligent early warning, and solve the shortcomings of the traditional static threshold method in dynamic environment. First, based on the type of flammable material (such as solid wood, paper), determine its flammable index (step S301), which is a quantitative index reflecting the flammable tendency of flammable material. The calculation of flammable index can refer to the above method (such as ignition point, self-ignition temperature, surface area / volume ratio), for example, the ignition point of wood is about 250°C, the self-ignition temperature is 400°C, and the flammable index may be 50 (0-100 range), while the paper may be 80 due to the lower ignition point (about 230°C). This index reflects the inherent risk of flammable material and provides a basis for subsequent calculation. Then, analyze the location characteristics of the flammable group and the high-heat block (step S302), and quantify the proximity of the two by edge mapping point distance (steps S3021-S3022). For example, in a virtual three-dimensional structure, the edge point coordinates of a piece of wood group are (4, 3, 1), and the edge point coordinates of the high-heat block are (4.5, 3, 1). The minimum edge distance is 0.5 meters, and the average distance is 1 meter. These distance data reflect the degree of direct threat of heat source to flammable material, and the closer the distance, the higher the risk.

[0048] In some embodiments disclosed by the present application, the method of visual analysis of the monitoring image comprises:

[0049] Step S201, using a visual analysis algorithm, identifying flammable materials in the monitoring image, and delimiting the identified flammable materials to form a flammable delimitation area.

[0050] Step S202, based on the shape and size of the flammable delimitation area, constructing a virtual flammable group.

[0051] In some embodiments disclosed by the present application, the method of constructing a virtual flammable group comprises:

[0052] Step S2021, determine the center point of the flammable material, and construct a vertical comparison surface at the center point, determine the flammable material mapping edge and the flammable material mapping area on the vertical comparison surface, uniformly set a plurality of shape analysis points on the flammable material mapping edge, determine the center shape analysis point between the shape analysis points, and calculate the analysis point relative distance of each shape analysis point relative to the center shape analysis point, and construct the analysis point relative distance group according to the original order, and identify the analysis point relative distance group as the shape characteristics of the flammable material, and identify the flammable material mapping area as the size characteristics of the flammable material.

[0053] In step S2022, the average relative distance and the relative distance variance of the analysis point relative distance in the analysis point relative distance group are calculated and analyzed, and the average relative distance of the combustible, the relative distance variance of the combustible, and the mapping area of the combustible are taken as the virtual combustible group construction conditions to construct the virtual combustible group.

[0054] In some embodiments of the present application, the method for constructing the virtual combustible group by taking the relative distance variance of the combustible and the mapping area of the combustible as the virtual combustible group construction conditions comprises the following steps:

[0055] In step S20221, the virtual three-dimensional scene structure is gridded, the planar mapping area corresponding to the over-center shape analysis point of the combustible delineation area is determined, and the grid corresponding to the planar mapping area is marked as a basic grid.

[0056] In step S20222, based on the average relative distance of the combustible, a number of grids above and below each basic grid are marked for expansion, so that the span of the grids in the same vertical direction falls within the preset distance ratio interval in proportion to the distance of the average relative distance, and the grids marked for expansion at this time are recorded as core grids.

[0057] In step S20223, based on the preset distance variance interval to which the relative distance variance belongs, the number of outer peripheral expansion layers is determined, and based on the number of outer peripheral expansion layers, the outer peripheral grids of the core grids are marked for expansion, and the grids marked for expansion at this time are recorded as warning grids.

[0058] In some embodiments of the present application, the method for configuring the high-heat block in the virtual three-dimensional scene structure comprises the following steps:

[0059] In step S203, the infrared thermal imaging image is analyzed for heat, and the continuous area in which the heat performance reaches the preset high-temperature requirement is delineated and recorded as a high-heat area.

[0060] In step S204, based on the area in which the high-heat area is mapped in the virtual three-dimensional scene structure, a high-heat spherical block is constructed in the virtual three-dimensional scene structure.

[0061] In some embodiments of the present application, the method for constructing the high-heat spherical block in the virtual three-dimensional scene structure comprises the following steps:

[0062] In step S2041, the area center point of the high-heat area is determined, and the vertical mapping area of the high-heat area relative to the vertical plane is determined, based on the preset area interval to which the vertical mapping area belongs, the radius of the high-heat spherical block is determined, and the area center point is taken as the center of the circle to construct the high-heat spherical block.

[0063] In some embodiments of the present disclosure, the method for determining the flammable risk of each flammable block based on the flammable characteristics of each flammable block and the location characteristics of the relative high-heat block comprises:

[0064] In step S301, the flammable index of the flammable material is determined based on the type of the flammable material, and the respective proximity block between the flammable block and the high-heat block is determined.

[0065] In step S302, the initial risk between the flammable block and the high-heat block is determined based on the location characteristics of the respective proximity block, and the flammable risk of the flammable block is determined in combination with the flammable index.

[0066] The method for determining the flammable index: According to the physical and chemical properties of the flammable material, the specific type is determined, such as wood, paper, plastic, etc. in solid flammable materials. Classification is carried out by observing the morphology (such as block, powder) and composition (such as cellulose, polymer), which can refer to the Material Safety Data Sheet (MSDS). For example, a piece of wood is identified as a cellulose-based solid. The basic properties of solid flammable materials are obtained, including the ignition point (the lowest sustained combustion temperature, such as wood about 250°C), the self-ignition temperature (the self-ignition temperature without a fire source, such as wood about 400°C), the surface area / volume ratio (powder is higher than block), the moisture content (affects the ignition point). These data can be obtained by experiment or database. For example, dry wood has low moisture content and high flammability. The influence of environmental conditions on flammability is evaluated, such as ambient temperature (close to the ignition point increases the risk), oxygen concentration (high concentration accelerates combustion), ventilation conditions (dry flammable material). For example, wood is more flammable in a high-temperature and ventilated workshop than in a humid and closed environment. According to the properties and environmental factors, a calculation formula is designed: flammable index = k1 × (1 / ignition point) + k2 × surface area / volume ratio + k3 × (1 / moisture content). The weights k1, k2, k3 are adjusted according to the type. For example, the wood has an ignition point of 250°C, a surface area / volume ratio of 0.1, and a moisture content of 5%, with k1 = 10000, k2 = 10, and k3 = 50, the calculated index is approximately 41. The model is verified by combustion experiment, and the ignition speed and heat release rate are measured to calibrate the weights. For example, the wood experiment shows moderate flammability, the index range is set to 0-100, and 41 belongs to moderate and low. The flammable index is output and classified: 0-20 low, 21-50 moderate, 51-80 high, and 81-100 extremely high. The wood index is 41, which belongs to moderate flammability, and the risk suggestion is attached.

[0067] In some embodiments of the present disclosure, the method for determining the flammable risk of each flammable block further comprises:

[0068] Step S3021, a plurality of edge mapping points are randomly set on the edges of the combustible block and the high-heat block respectively, and one edge mapping point is randomly selected on the edges of the combustible block and the high-heat block respectively to combine, to obtain a plurality of edge mapping point groups.

[0069] Step S3022, the edge mapping point distance between the edge mapping points in each edge mapping point group is determined, and the minimum edge mapping point distance and the average mapping point distance are determined, and based on the average mapping point distance and the minimum edge mapping point distance, the initial risk of the combustible block and the high-heat block is determined, and the combustible risk of the combustible block is determined in combination with the combustible index.

[0070] Wherein, the expression for calculating the combustible risk is:

[0071] .

[0072] Wherein, F is the combustible risk, K is the combustible index, is a preset reference average mapping point distance, is the average mapping point distance, is a preset reference minimum edge mapping point distance, is the minimum edge mapping point distance, L is the minimum edge mapping point distance influence adjustment coefficient, and b is the minimum edge mapping point distance influence adjustment constant.

[0073] Some embodiments disclosed in the present application also disclose a distributed fire warning system based on Internet of Things and artificial intelligence, comprising:

[0074] The first module is used for determining the site structure design of the fire monitoring site, and constructing a virtual three-dimensional site structure based on the site structure design;

[0075] The second module is used for acquiring an infrared thermal imaging image and a monitoring image of the site, performing visual analysis on the monitoring image to determine combustible materials in the site, constructing a virtual combustible block based on the size feature and the shape feature of the combustible materials in the site, and configuring the virtual combustible block in the virtual three-dimensional site structure, and configuring a high-heat block in the virtual three-dimensional site structure based on the heat distribution feature of the infrared thermal imaging image.

[0076] The third module is used for determining the combustible risk of each combustible block based on the combustible feature of each combustible block and the position feature of the relative high-heat block, and if the combustible risk is greater than or equal to a threshold value, a pre-alarm is performed.

[0077] The application discloses a distributed fire starting early warning method and system based on an Internet of Things and artificial intelligence, relates to the technical field of fire starting early warning, and determines the site structure design of a fire starting monitoring site, and constructs a virtual three-dimensional site structure; infrared thermal imaging images and monitoring images are used in combination with visual analysis technology to identify site combustibles, virtual combustible blocks are constructed based on the size and shape features of the combustibles, and are configured in the virtual three-dimensional site structure, and a high heat block is configured according to heat distribution features; the combustible features of each combustible block and the positional relationship with the high heat block are comprehensively considered to calculate a combustible risk, and early warning is triggered when the risk value is greater than or equal to a threshold value; the application realizes distributed fire starting risk assessment by collecting multi-source data in real time through the Internet of Things and accurately analyzing the spatial relationship between combustibles and heat sources through artificial intelligence.

[0078] Through the description of the above implementation mode, those skilled in the art can clearly understand that the application can be implemented by hardware, or can be implemented by means of software and a necessary general hardware platform. Based on such understanding, the technical solutions of the application can be embodied in the form of a software product, which can be stored in a nonvolatile storage medium (which can be a CD-ROM, a U disk, a mobile hard disk, etc.), and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the method described in each implementation scenario of the application.

[0079] Finally, it should be noted that: the above examples are only used to illustrate the technical solutions of the application rather than limit them, and although the application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that: it can still modify or equivalently replace the technical solutions of the application, and these modifications or equivalent replacements cannot make the modified technical solutions deviate from the spirit and scope of the technical solutions of the application.

Claims

1. A distributed fire warning method based on Internet of Things and artificial intelligence, characterized in that, The method comprises the following steps: Step S100, determining a site structure design of a fire monitoring site, and constructing a virtual three-dimensional site structure based on the site structure design; Step S200, acquiring an infrared thermal imaging image and a monitoring image of the site, performing visual analysis on the monitoring image to determine combustible materials in the site, constructing a virtual combustible material block based on size features and shape features of the combustible materials in the site, and configuring the virtual combustible material block in the virtual three-dimensional site structure, and performing high-heat block configuration on the virtual three-dimensional site structure based on heat distribution features of the infrared thermal imaging image; Step S300, determining a combustible risk of each combustible material block based on combustible features of each combustible material block and position features of the relative high-heat block, and performing a pre-alarm if the combustible risk is greater than or equal to a threshold value; The method for determining the combustible risk of each combustible material block based on the combustible features of each combustible material block and the position features of the relative high-heat block comprises the following steps: Step S301, determining a combustible index of the combustible material based on a type to which the combustible material belongs, and determining respective proximity blocks between the combustible block and the high-heat block; Step S302, determining initial risks of the combustible block and the high-heat block based on position features between the respective proximity blocks, and determining the combustible risk of the combustible material block in combination with the combustible index; The method for determining the combustible risk of the combustible material block further comprises the following steps: Step S3021, randomly setting a plurality of edge mapping points on edges of the combustible block and the high-heat block respectively, and randomly selecting one edge mapping point on the edges of the combustible block and the high-heat block respectively to combine, to obtain a plurality of edge mapping point groups; Step S3022, determining edge mapping point distances between the edge mapping points in each edge mapping point group, determining a minimum edge mapping point distance and an average mapping point distance, and determining the initial risks of the combustible block and the high-heat block based on the average mapping point distance and the minimum edge mapping point distance, and determining the combustible risk of the combustible material block in combination with the combustible index; Wherein, an expression for calculating the combustible risk is as follows: ； Wherein, F is the flammable risk, K is the flammable index, is a preset reference average mapping point distance, is an average mapping point distance, is a preset reference minimum edge mapping point distance, is a minimum edge mapping point distance, L is a minimum edge mapping point distance influence adjustment coefficient, and b is a minimum edge mapping point distance influence adjustment constant. 2.The distributed fire warning method based on the Internet of Things and artificial intelligence according to claim 1, wherein, The method for performing visual analysis on the monitoring image comprises the following steps: Step S201, using a visual analysis algorithm to identify the combustible material in the monitoring image, and performing edge demarcation on the identified combustible material to form a combustible material demarcated area; Step S202, constructing a virtual combustible material block based on the shape and size of the combustible material demarcated area. 3.The distributed fire warning method based on the Internet of Things and artificial intelligence according to claim 2, wherein, The method for constructing the virtual combustible material block comprises the following steps: Step S2021, determining a combustible material center point, constructing a vertical comparison surface at the center point, determining a combustible material mapping edge and a combustible material mapping area on the vertical comparison surface, uniformly setting a plurality of shape analysis points on the combustible material mapping edge, determining a center shape analysis point between the shape analysis points, calculating an analysis point relative distance of each shape analysis point relative to the center shape analysis point, constructing an analysis point relative distance group in the original order, identifying the analysis point relative distance group as the shape features of the combustible material, and identifying the combustible material mapping area as the size features of the combustible material; In step S2022, the average relative distance and the relative distance variance of the analysis point relative distance in the analysis point relative distance group are calculated and analyzed, and the average relative distance of the combustible, the relative distance variance of the combustible, and the mapping area of the combustible are taken as the virtual combustible group construction conditions to construct the virtual combustible group. 4.The distributed fire warning method based on the Internet of Things and artificial intelligence according to claim 3, wherein, The method for constructing the virtual combustible group by taking the relative distance variance of the combustible and the mapping area of the combustible as the virtual combustible group construction conditions comprises the following steps: In step S20221, the virtual three-dimensional site structure is gridded, the planar mapping area corresponding to the over-center shape analysis point of the combustible delineation area is determined, and the grid corresponding to the planar mapping area is marked as a basic grid; In step S20222, based on the average relative distance of the combustible, a number of grids above and below each basic grid are marked for expansion, so that the span of the grids in the same vertical direction falls within the preset distance ratio interval in proportion to the distance of the average relative distance, and the grids marked for expansion at this time are recorded as core grids; In step S20223, based on the preset distance variance interval to which the relative distance variance belongs, the number of outer peripheral expansion layers is determined, and the outer peripheral grids of the core grids are marked for expansion based on the number of outer peripheral expansion layers, and the grids marked for expansion at this time are recorded as warning grids. 5.The distributed fire warning method based on the Internet of Things and artificial intelligence according to claim 1, wherein, The method for configuring the high-heat block in the virtual three-dimensional site structure comprises the following steps: In step S203, the infrared thermal imaging image is analyzed for heat, and a continuous area in which the heat performance reaches a preset high-temperature requirement is delineated and recorded as a high-heat area; In step S204, based on the area in which the high-heat area is mapped in the virtual three-dimensional site structure, a high-heat spherical block is constructed in the virtual three-dimensional site structure. 6.The distributed fire warning method based on the Internet of Things and artificial intelligence according to claim 5, wherein, The method for constructing the high-heat spherical block in the virtual three-dimensional site structure comprises the following steps: In step S2041, the area center point of the high-heat area is determined, and the vertical mapping area of the high-heat area relative to the vertical plane is determined, the radius of the high-heat spherical block is determined based on the preset area interval to which the vertical mapping area belongs, and the area center point is taken as the center of a circle to construct the high-heat spherical block.

7. A distributed fire warning system based on Internet of Things and artificial intelligence, characterized in that, The method for executing the distributed hot work pre-warning method in any one of claims 1-6 comprises: A first module is configured to determine the site structure design of the hot work monitoring site, and construct a virtual three-dimensional site structure based on the site structure design; A second module is configured to acquire the infrared thermal imaging image and the monitoring image of the site, perform visual analysis on the monitoring image to determine the combustible in the site, construct a virtual combustible group based on the size feature and the shape feature of the combustible in the site, and configure the virtual combustible group in the virtual three-dimensional site structure, and perform high-heat block configuration on the virtual three-dimensional site structure based on the heat distribution feature of the infrared thermal imaging image; A third module is configured to determine the combustible risk of each combustible group based on the combustible feature of each combustible group and the position feature relative to the high-heat block, and perform pre-warning if the combustible risk is greater than or equal to a threshold value.

Images