Intelligent fire early warning system based on multi-source data

Abstract

The application discloses a fire intelligent early warning system based on multi-source data, relates to the technical field of fire early warning, and solves the technical problem that the existing technology cannot guarantee the fire spreading precision and cannot be applied to various fire early warning scenes because of the single influencing factor considered in fire spreading prediction; the application acquires basic data through a plurality of edge collection modules, acquires target data corresponding to a target region after data processing of the basic data, and acquires a fire spreading speed by combining the target data with an artificial intelligence model; the fire spreading is predicted by comprehensively considering various influencing factors, the fire spreading precision is guaranteed, and resource scheduling can be performed in time; in the process of acquiring the fire spreading speed through the speed prediction model, the target region is divided into a plurality of sub-regions with different flammability degrees according to regional data, the fire spreading speed is predicted by taking the sub-regions as basic units, the precision of the fire spreading speed is guaranteed, and the accuracy of the fire spreading prediction is improved.

Description

Technical Field

[0001] This invention belongs to the field of fire early warning and relates to intelligent fire early warning technology based on multi-source data, specifically an intelligent fire early warning system based on multi-source data. Background Technology

[0002] Fires are extremely destructive and difficult to control. They not only cause property damage and air pollution, but also have various impacts on people's health. Various measures have been tried to address fire early warning systems, but none have been able to meet the need for intelligent early warning systems that are available in all times and all spaces.

[0003] Existing technology (patent application CN113139272A) discloses a method, device, equipment, and storage medium for predicting forest fire spread. It inputs current fire data into a forest fire spread rate model constructed based on a cellular automata algorithm to obtain the predicted forest fire spread results, thus improving the accuracy of the prediction. However, this existing technology is mainly applied to forest fire spread prediction, relying solely on fire data and considering only a single influencing factor. This results in inconsistent accuracy and makes it unsuitable for various fire warning scenarios. Therefore, there is an urgent need for an intelligent fire warning system based on multi-source data. Summary of the Invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a fire intelligent early warning system based on multi-source data to solve the technical problem that the existing technology considers only one influencing factor in fire spread prediction, which leads to the inability to guarantee the accuracy of fire spread and to be applied to various fire early warning scenarios.

[0005] To achieve the above objectives, the first aspect of the present invention provides a fire intelligent early warning system based on multi-source data, including a central analysis module, a data display module and several edge analysis modules connected thereto, wherein the edge analysis modules are connected to a data acquisition device;

[0006] Edge analysis module: In the event of a fire, it collects basic data through connected data acquisition devices; this basic data includes environmental data and area condition data; and

[0007] The basic data is processed to obtain raw data, which is then sent to the central analysis module; the data processing includes data filtering and data integration.

[0008] Central analysis module: Identifies the target area based on the aforementioned area data, and divides it into several sub-regions according to the identification results; among which, the target area includes the fire area; and

[0009] The fire spread rate is obtained for several sub-regions based on raw data and artificial intelligence models; the fire spread process is simulated based on the fire spread rate, and the data is visualized through the data display module.

[0010] Preferably, the central analysis module is communicatively and / or electrically connected to the data display module and several edge analysis modules, and the data display module is used for data display.

[0011] The edge analysis module is configured according to configuration rules and is connected to the data acquisition device; the configuration rules include region-based configuration or data type-based configuration.

[0012] Preferably, before the edge analysis module collects basic data, the central analysis module generates a data acquisition signal and sends it to the edge analysis module, including:

[0013] Receive multi-source data; wherein, the multi-source data includes video data sent by smart terminals and watchtowers, and smart terminals include mobile phones and computers;

[0014] The video data is identified and analyzed using image analysis technology. The fire level is marked according to the analysis results, and it is determined whether to generate a data acquisition signal based on the fire level.

[0015] When the fire level is greater than the level threshold, a data acquisition signal is generated and sent to several edge analysis modules; wherein the level threshold is set based on experience.

[0016] Preferably, the edge analysis module performs data processing on the basic data to obtain the raw data, including:

[0017] The environmental data is extracted from the basic data, and an environmental distribution image is generated based on the environmental data and the configuration rules; wherein, the environmental data includes wind force, wind direction, temperature or humidity;

[0018] Extract the area condition data from the basic data, and render and generate an area condition distribution image based on the area condition data and the configuration rules; wherein, the area condition data includes DEM data and material type data;

[0019] The environmental distribution image and the area condition distribution image are matched to obtain the raw data.

[0020] Preferably, the central analysis module identifies the material type in the target area and divides the target area into several sub-regions based on the identification result, including:

[0021] Obtain the area condition data corresponding to the target area; wherein, the target area is generated by expanding the area according to a set area based on the location of the fire;

[0022] Identify several material types in the area condition data, and divide the target area into several sub-areas based on the several material types; wherein the material types or flammability ratings of the sub-areas are the same.

[0023] Preferably, the central analysis module determines the fire spread rate of several sub-regions based on raw data, including:

[0024] Based on the target region, the original data is spliced ​​or overlaid to obtain the target data;

[0025] Extract model input data corresponding to the sub-region from the target data; wherein, the model input data includes environmental data and area condition data of the corresponding sub-region;

[0026] The input data of the model is input into the speed prediction model to obtain the output fire spread speed; wherein, the speed prediction model is constructed based on an artificial intelligence model.

[0027] Preferably, the speed prediction model is trained by combining the artificial intelligence model with standard training data, including:

[0028] Obtain standard training data; where the standard training data consists of the fire spread rate of various material types in different fire simulation environments;

[0029] The standard training data is expanded using interpolation, and the expanded data is used to train the constructed artificial intelligence model to obtain the speed prediction model.

[0030] Preferably, the central analysis module simulates and obtains the fire spread process of the target area based on the fire spread rate of several sub-regions, and displays the fire spread process through the data display module; and

[0031] The corresponding fire spread time is calculated based on the set boundary, or the corresponding fire spread area is calculated based on the set time; resources are allocated according to the fire spread time or fire spread area.

[0032] Compared with the prior art, the beneficial effects of the present invention are:

[0033] 1. This invention acquires basic data through several edge acquisition modules, processes the basic data to obtain target data corresponding to the target area, and combines the target data with an artificial intelligence model to obtain the fire spread rate; it comprehensively considers multiple influencing factors to achieve fire spread prediction, broadens the application scenarios while ensuring the accuracy of fire spread, and facilitates timely resource scheduling.

[0034] 2. In the process of obtaining the fire spread rate through the speed prediction model, the present invention divides the target area into several sub-areas with different flammability based on the area condition data, and uses each sub-area as the basic unit to predict the fire spread rate, thereby ensuring the accuracy of the fire spread rate and improving the accuracy of fire spread prediction. Attached Figure Description

[0035] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0036] Figure 1 This is a schematic diagram of the working steps of the present invention;

[0037] Figure 2 This is a schematic diagram of the system principle of the present invention. Detailed Implementation

[0038] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. 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.

[0039] Please see Figures 1-2 The first aspect of the present invention provides a fire intelligent early warning system based on multi-source data, including a central analysis module, a data display module and several edge analysis modules connected thereto, and the edge analysis modules are connected to the data acquisition device;

[0040] Edge analysis module: When a fire occurs, it collects basic data through connected data acquisition devices; it also processes the basic data to obtain raw data and sends it to the central analysis module.

[0041] Central Analysis Module: Identifies the target area by combining regional data and divides it into several sub-regions based on the identification results; obtains the fire spread rate corresponding to several sub-regions based on raw data and artificial intelligence models; simulates the fire spread process based on the fire spread rate and visualizes it through the data display module.

[0042] This invention application primarily addresses the problem in existing technologies where the reference influencing factors is singular, resulting in a limited range of applicable scenarios. The central analysis module is a comprehensive processor, mainly performing data processing; the edge analysis module mainly collects data according to the set tasks, while also possessing data processing capabilities.

[0043] The basic data includes environmental data and area data. Environmental data mainly refers to variable data such as wind force, wind direction, temperature, and humidity that can affect the fire's trajectory. Area data consists of fixed data such as geographical data, building data, and vegetation data corresponding to the fire area that influence the fire's trajectory. Combining environmental and area data ensures the accuracy of the obtained fire spread rate.

[0044] In this invention application, the central analysis module is communicatively and / or electrically connected to the data display module and several edge analysis modules, and the data display module is used for data display; the edge analysis modules are configured according to configuration rules, and the edge analysis modules are connected to the data acquisition device; wherein, the configuration rules include configuration based on region or configuration based on data type.

[0045] It is important to understand that the edge analysis module needs to be configured according to the configuration rules. When the configuration rule is based on region, the number of edge analysis modules is determined based on the data processing capability of the edge analysis module and the size of the target region to ensure data processing efficiency. When the configuration rule is based on data type, the types of basic data are first determined, and then at least one edge analysis module is configured for each type of basic data to reduce the switching of data processing algorithms.

[0046] In this invention application, before the edge analysis module collects basic data, a data acquisition signal is generated by the central analysis module and sent to the edge analysis module, including:

[0047] It receives multi-source data; it performs identification and analysis on video data based on image analysis technology, marks the fire level according to the analysis results, and determines whether to generate a data acquisition signal based on the fire level; when the fire level is greater than the level threshold, it generates a data acquisition signal and sends the data acquisition signal to several edge analysis modules.

[0048] Multi-source data includes video data transmitted from smart terminals and watchtowers. Smart terminals include mobile phones and computers; video data can be acquired through fixed equipment such as watchtowers and surveillance cameras, or through personal smart terminals. Image recognition technology is used to identify the size of the fire in the video data, thereby determining the fire severity level. When the fire is small, there is no need for fire spread analysis. For example, if a quilt drying in an open area catches fire, it is easily dealt with promptly and generally will not spread. However, if a new energy vehicle spontaneously combusts in a parking lot, the fire could easily spread.

[0049] When the fire level exceeds the threshold, it indicates that the fire is large and easily spreads. Therefore, data acquisition signals are generated to analyze the fire spread. At the same time, fire-fighting resources around the fire should be allocated to reduce losses.

[0050] The edge analysis module in this invention application processes the basic data to obtain the raw data, including:

[0051] Environmental data is extracted from the basic data, and an environmental distribution image is generated based on the environmental data and configuration rules. Regional condition data is extracted from the basic data, and a regional condition distribution image is generated based on the regional condition data and configuration rules. The environmental distribution image and the regional condition distribution image are matched to obtain the original data.

[0052] Because the configuration rules for edge analysis modules differ, the content of the basic data also varies. When the configuration rule is set to region, the basic data obtained by the edge analysis module includes various types of environmental data and various types of area condition data for the corresponding region. Direct rendering and matching are then performed to obtain the original data for that region. Concatenating the original data from each region yields the original data corresponding to the target region. However, when the configuration rule is set to data type, the data type obtained by the edge analysis module is either a single type of environmental data or area condition data. In this case, the original data obtained by processing the data from several edge analysis modules refers to the data corresponding to the target region, i.e., the sum of the original data from several regions.

[0053] It should be understood that the area corresponding to the edge analysis module is not the same as the sub-region in this invention application. The area corresponding to the edge analysis module is to improve the efficiency of data acquisition and processing, while the sub-region is to analyze the speed of fire spread.

[0054] In this invention application, the central analysis module identifies the material type in the target region and divides the target region into several sub-regions based on the identification results, including:

[0055] Obtain the zonal data corresponding to the target area; identify several material types in the zonal data, and divide the target area into several sub-regions based on the material types.

[0056] To ensure the accuracy of the obtained fire spread rate, analysis is needed for different material types. Under the same conditions, flammable materials have a greater fire spread rate than non-flammable materials. This invention application extracts the material type of the target area based on area condition data, and then divides it into several sub-areas. It should be understood that the material type referred to here is a generalization based on objects in the area condition data. For example, residential buildings and roads can be classified as one type of material, while vegetation and turf can be classified as another. In the absence of classification, several sub-areas can also be divided solely based on flammability, that is, the material types or flammability levels in the sub-areas are consistent.

[0057] The target area is generated by expanding the area according to the location of the fire. For example, if a fire occurs at a certain point, a circular or rectangular area of ​​a certain size is defined as the target area, centered on that point. This can also be understood as the area that the fire may affect.

[0058] The area data includes DEM data and material type data. The DEM data is obtained through a third-party database, while the material type data is obtained by summarizing the objects on the surface of the target area. As mentioned above, summarizing the material types of surface objects is not necessary; they can also be classified directly based on their flammability.

[0059] The central analysis module in this invention application determines the fire spread rate of several sub-regions based on raw data, including:

[0060] Based on the target area, the original data is spliced ​​or overlaid to obtain the target data; the model input data corresponding to the sub-region is extracted from the target data; the model input data is input into the velocity prediction model to obtain the output fire spread velocity.

[0061] The raw data obtained through the edge analysis module is fragmented. If each edge analysis module is responsible for the basic data collection of a region, the raw data corresponding to several edge analysis modules needs to be stitched together to form the target data corresponding to the target region. If each edge analysis module is responsible for the basic data collection of a certain type, the raw data corresponding to several edge analysis modules needs to be overlaid (similar to overlay processing in the field of remote sensing image processing) to obtain the target data corresponding to the target region.

[0062] The model input data includes environmental and area condition data for the corresponding sub-region, namely wind force, wind direction (which can be replaced by numerical labels), temperature, humidity, average DEM value, flammability rating, etc. By inputting the model input data into the velocity prediction model, the fire spread velocity corresponding to the sub-region can be obtained.

[0063] This invention application combines an artificial intelligence model with standard training data to train and obtain a speed prediction model, including:

[0064] Obtain standard training data; expand the standard training data using interpolation; train the constructed artificial intelligence model using the expanded data to obtain a speed prediction model.

[0065] The standard training data consists of fire spread rates for various material types in different fire simulation environments. The differences in these simulation environments mainly lie in the environmental data and DEM data. Before training, to ensure the usability of the velocity prediction model, it is also necessary to ensure the total amount of standard training data. This can be achieved by interpolating the standard training data using interpolation methods to obtain sufficient standard training data.

[0066] Artificial intelligence models include deep convolutional neural network models or RBF neural network models. During the training process, the artificial intelligence model is improved by analyzing the relationship between the fire simulation environment, various types of materials (various flammability levels) and the fire spread rate in the standard training data.

[0067] In this invention application, the central analysis module simulates and obtains the fire spread process of the target area based on the fire spread speed of several sub-regions, and displays the fire spread process through the data display module; it also calculates and obtains the corresponding fire spread time based on a set boundary, or calculates and obtains the corresponding fire spread area based on a set time; and performs resource scheduling based on the fire spread time or fire spread area.

[0068] After obtaining the fire spread rate in each sub-region, the fire spread process in the target area can be simulated. Throughout the simulation, resources can be flexibly allocated based on set boundaries or set timeframes. It's important to understand that, under boundary conditions, the time required to reach the set boundary can be calculated based on the fire spread rate; under time-based conditions, the area the fire spreads within the set timeframe can be calculated based on the fire spread rate. Therefore, firefighting resources can be flexibly allocated based on either time or area to reduce fire losses.

[0069] Working principle of the invention:

[0070] When a fire occurs, the edge analysis module collects basic data through the data acquisition device connected to it; it processes the basic data to obtain raw data and sends it to the central analysis module.

[0071] The central analysis module combines the regional data in the original data to identify the target area and obtain several sub-regions; based on the original data and the artificial intelligence model, it obtains the fire spread rate corresponding to the sub-regions; based on the fire spread rate, it simulates the fire spread process and displays it visually.

[0072] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A fire early warning system based on multi-source data, comprising a central analysis module, a data display module connected thereto, and several edge analysis modules, wherein the edge analysis modules are connected to data acquisition equipment, characterized in that: Edge analysis module: In the event of a fire, it collects basic data through connected data acquisition devices; this basic data includes environmental data and area condition data; and The basic data is processed to obtain raw data, which is then sent to the central analysis module; the data processing includes data filtering and data integration. Central analysis module: Identifies the target area based on the aforementioned area data, and divides it into several sub-regions according to the identification results; among which, the target area includes the fire area; and The fire spread rate is obtained for several sub-regions based on raw data and artificial intelligence models; the fire spread process is simulated based on the fire spread rate and visualized through the data display module; The central analysis module identifies the material type in the target area and divides the target area into several sub-regions based on the identification results, including: Obtain the area condition data corresponding to the target area; wherein, the target area is generated by expanding the area according to a set area based on the location of the fire; Identify several material types in the area condition data, and divide the target area into several sub-areas based on the several material types; wherein the material types or flammability ratings are the same in the sub-areas. The central analysis module determines the fire spread rate in several sub-regions based on raw data, including: Based on the target region, the original data is spliced ​​or overlaid to obtain the target data; Extract model input data corresponding to the sub-region from the target data; wherein, the model input data includes environmental data and area condition data of the corresponding sub-region; The input data of the model is input into the speed prediction model to obtain the output fire spread speed; wherein, the speed prediction model is constructed based on an artificial intelligence model.

2. The intelligent fire early warning system based on multi-source data according to claim 1, characterized in that, The central analysis module is communicatively and / or electrically connected to the data display module and several edge analysis modules, and the data display module is used for data display. The edge analysis module is configured according to configuration rules and is connected to the data acquisition device; the configuration rules include region-based configuration or data type-based configuration.

3. The intelligent fire early warning system based on multi-source data according to claim 2, characterized in that, Before the edge analysis module collects basic data, the central analysis module generates a data acquisition signal and sends it to the edge analysis module, including: Receive multi-source data; wherein, the multi-source data includes video data sent by smart terminals and watchtowers, and smart terminals include mobile phones and computers; The video data is identified and analyzed using image analysis technology. The fire level is marked according to the analysis results, and it is determined whether to generate a data acquisition signal based on the fire level. When the fire level is greater than the level threshold, a data acquisition signal is generated and sent to several edge analysis modules; wherein the level threshold is set based on experience.

4. The intelligent fire early warning system based on multi-source data according to claim 3, characterized in that, The edge analysis module processes the basic data to obtain the raw data, including: The environmental data is extracted from the basic data, and an environmental distribution image is generated based on the environmental data and the configuration rules; wherein, the environmental data includes wind force, wind direction, temperature or humidity; Extract the area condition data from the basic data, and render and generate an area condition distribution image based on the area condition data and the configuration rules; wherein, the area condition data includes DEM data and material type data; The environmental distribution image and the area condition distribution image are matched to obtain the raw data.

5. The intelligent fire early warning system based on multi-source data according to claim 1, characterized in that, The speed prediction model is trained by combining the artificial intelligence model with standard training data, including: Obtain standard training data; where the standard training data consists of the fire spread rate of various material types in different fire simulation environments; The standard training data is expanded using interpolation, and the expanded data is used to train the constructed artificial intelligence model to obtain the speed prediction model.

6. The intelligent fire early warning system based on multi-source data according to claim 5, characterized in that, The central analysis module simulates and obtains the fire spread process of the target area based on the fire spread speed of several sub-regions, and displays the fire spread process through the data display module; as well as The corresponding fire spread time is calculated based on the set boundary, or the corresponding fire spread area is calculated based on the set time; resources are allocated according to the fire spread time or fire spread area.

Images