Fire point fire control display method and system based on digital twinning

Abstract

The invention provides a fire point fire control display method and system based on digital twinning, and relates to the technical field of digital twinning fire analysis, and the method comprises the steps: obtaining image data in a garage, inputting the preprocessed image data into a fire recognition model, and obtaining a fire recognition result; acquiring fire data, inputting the fire data into the situation development prediction model, and acquiring a fire behavior development prediction result; based on the fire behavior development prediction result, utilizing a support vector machine algorithm to identify a potential fire risk and a risk level in the scene, and generating a scene risk assessment result; based on a scene risk assessment result, constructing a fire-fighting relation graph, and analyzing a dynamic relation between entities through a graph attention network to generate a dynamic relation graph; simulating a scene and acquiring real-time data change by utilizing a data visualization technology based on the dynamic relation graph, and constructing a digital twinborn model to generate a visual interface; fire-fighting actions are deployed and controlled according to the visual interface, and accurate fire extinguishing in the garage is achieved.

Description

Technical Field

[0001] The present disclosure relates to the technical field of digital twin fire analysis, and in particular to a fire point fire control display method and system based on digital twin. Background Art

[0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art.

[0003] Digital Twin refers to a virtual model that is highly consistent with the physical entity created in a virtual space through digital means. Through data connection, the two are kept dynamically synchronized throughout the entire life cycle, thereby achieving real-time monitoring, simulation and optimization of the physical entity.

[0004] Digital twin technology usually consists of physical entities, virtual models and data connections. Physical entities are the basis of digital twin technology, referring to devices, systems or processes in the real world. Physical entities transmit their own operating status, environmental conditions and other information to the digital twin system in real time through sensors and data acquisition devices. Virtual models are the mapping of physical entities in digital space, and usually use a variety of modeling techniques, such as geometric modeling, physical modeling and behavioral modeling. Data connection is a bridge between physical entities and virtual models, including not only the transmission of state data from physical entities to virtual models, but also control instructions and optimization suggestions fed back from virtual models to physical entities.

[0005] In the complex scene of garage fire fighting and rescue, the application of digital twin technology can realize important emergency rescue functions such as real-time monitoring of the fire scene, simulation of fire development, and optimization of rescue plans, greatly improving the scientificity and effectiveness of fire fighting and rescue. Since the identification of fire conditions and the timeliness and accuracy of monitoring are the key to fire fighting control and rescue, the simulation of complex fire scenes and the dynamic development prediction perception are extremely important for fire control and real-time control display. In the complex scene of the garage, since there are many types of vehicles stored in the scene, including electric cars or oil cars, it is necessary to simulate this complex scene in reality to avoid greater losses.

[0006] However, when dealing with complex fire scenes in garages, the current fire simulation and prediction is difficult due to the dim environment in the garage. Object recognition in a dim environment is relatively difficult and lacks real-time and accuracy. Existing scene modeling methods are mostly static, but the flow of vehicles and personnel in the garage is dynamic. Existing modeling methods cannot meet the dynamic modeling requirements and cannot quickly simulate and locate fire scenes. Summary of the invention

[0007] In order to solve the above problems, the present invention proposes a fire control display method and system for fire starting points based on digital twins. By collecting and analyzing sensor data through digital twins, early identification and warning of fires can be achieved. With the help of fire simulation and situation awareness functions, a regression variable selection-neural network model is constructed to predict the development of fires, improve the accuracy of fire development prediction at the fire starting point, achieve precise fire control, and perform real-time fire screen visualization, significantly optimizing the whole process management of fire fighting and rescue.

[0008] According to some embodiments, the present disclosure adopts the following technical solutions: The fire control display method of the fire point based on digital twin includes: Acquire image data in the garage, and pre-process the image data in the garage; The pre-processed image data is input into the fire recognition model to obtain the fire recognition result and locate the fire point; Obtaining fire data of the fire starting point, inputting the fire data into a situation development prediction model, and obtaining a fire development prediction result; based on the fire development prediction result, using a support vector machine algorithm to identify potential fire risks and risk levels in the scene, and generating a scene risk assessment result; Based on the results of scenario risk assessment, a firefighting relationship diagram between firefighting equipment, vehicles and fire points is constructed, and the dynamic relationship between entities is analyzed through the graph attention network to generate a dynamic relationship diagram; Based on the dynamic relationship diagram, using data visualization technology, simulate the scene and obtain real-time data changes, build a digital twin model to update and dynamically display data in real time, and generate a visualization interface; Firefighting actions are deployed and controlled based on the visual interface to achieve precise fire extinguishing in the garage.

[0009] According to some embodiments, the present disclosure adopts the following technical solutions: The fire control and display system based on digital twins includes: A data acquisition module, used for acquiring image data in the garage and preprocessing the image data in the garage; The fire identification module is used to input the pre-processed image data into the fire identification model to obtain the fire identification result and locate the fire point; An assessment module is used to obtain fire data of the fire point, input the fire data into the situation development prediction model, and obtain the fire development prediction result; based on the fire development prediction result, a support vector machine algorithm is used to identify the potential fire risk and risk level in the scene, and generate a scene risk assessment result; The model building and visualization module is used to build a firefighting relationship diagram between firefighting equipment, vehicles and fire points based on the results of scenario risk assessment, analyze the dynamic relationship between entities through the graph attention network, and generate a dynamic relationship diagram; based on the dynamic relationship diagram, use data visualization technology to simulate the scenario and obtain real-time data changes, build a digital twin model for real-time update and dynamic display of data, and generate a visualization interface; The fire control module is used to deploy and control firefighting actions based on a visual interface to achieve precise fire extinguishing in the garage.

[0010] According to some embodiments, the present disclosure adopts the following technical solutions: A computer program product includes a computer program, which, when executed by a processor, implements the fire control display method for a fire point based on digital twins.

[0011] According to some embodiments, the present disclosure adopts the following technical solutions: A non-transitory computer-readable storage medium, which is used to store computer instructions. When the computer instructions are executed by a processor, the fire control display method of the fire point based on digital twin is implemented.

[0012] According to some embodiments, the present disclosure adopts the following technical solutions: An electronic device comprises: a processor, a memory and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory so that the electronic device executes the fire control display method based on digital twin.

[0013] Compared with the prior art, the present invention has the following beneficial effects: The fire control display method of the fire starting point based on digital twin disclosed in the present invention identifies the fire starting point through a fire recognition model based on the YOLOv5s model. The fire recognition model consists of three parts: Backbone network, Neck network and Prediction. The Backbone network of the fire recognition model uses deep separable convolution, and an attention mechanism that pays attention to the number of channels is added to the bottom layer of the Backbone network. A parameter-free attention mechanism is introduced into the Neck network. The parameter-free attention mechanism is composed of an activation function and an energy function connected in sequence, which has advantages in fire recognition accuracy and detection speed, reduces redundancy in the fire recognition model, and achieves a balance between detection speed and accuracy; while reducing the parameters and inference time of the fire recognition network, the accuracy of fire recognition is guaranteed, so that the fire recognition network improves the ability to recognize small target fire smoke and flames without increasing parameters.

[0014] The fire control display method of the fire point based on digital twin disclosed in the present invention adopts data visualization technology to simulate the fire scene and analyze the real-time data changes, uses three-dimensional modeling technology to build a digital twin model of the fire scene, and uses real-time dynamic data processing methods to perform real-time update and dynamic display of data to generate a comprehensive visualization interface, which significantly improves the response capability of the fire control room to complex situations, can effectively manage and analyze fire data, and through accurate fire development prediction and risk identification, provides intuitive and real-time data support for command decision-making through three-dimensional visualization technology, significantly improving rescue efficiency and the scientific nature of command decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The accompanying drawings constituting a part of the present disclosure are used to provide a further understanding of the present disclosure. The illustrative embodiments of the present disclosure and their descriptions are used to explain the present disclosure and do not constitute an improper limitation on the present disclosure.

[0016] Figure 1 A flow chart of a fire point fire control display method according to an embodiment of the present disclosure; DETAILED DESCRIPTION The present disclosure is further described below in conjunction with the accompanying drawings and embodiments.

[0017] It should be noted that the following detailed descriptions are all illustrative and are intended to provide further explanation of the present disclosure. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present disclosure belongs.

[0018] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present disclosure. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0019] Example 1 In one embodiment of the present disclosure, a fire control display method based on digital twin is provided, including: Step 1: Obtain image data in the garage and pre-process the image data in the garage; Step 2: Input the preprocessed image data into the fire recognition model to obtain the fire recognition result and locate the fire point; Step 3: Obtain fire data of the fire starting point, input the fire data into the situation development prediction model, and obtain the fire development prediction result; based on the fire development prediction result, use the support vector machine algorithm to identify the potential fire risk and risk level in the scene, and generate the scene risk assessment result; Step 4: Based on the scenario risk assessment results, a firefighting relationship diagram between firefighting equipment, vehicles and fire points is constructed, and the dynamic relationship between entities is analyzed through the graph attention network to generate a dynamic relationship diagram; Step 5: Based on the dynamic relationship diagram, use data visualization technology to simulate the scene and obtain real-time data changes, build a digital twin model to update and dynamically display data in real time, and generate a visualization interface; Step 6: Deploy and control firefighting actions based on the visual interface to achieve precise fire extinguishing in the garage.

[0020] As an embodiment, the specific implementation process of the fire point fire control display method based on digital twin disclosed in the present invention includes the following steps: Step 1: Obtain the image data in the garage and pre-process the image data in the garage; Specifically, by shooting a video in the garage, the shot video is processed into frame image data, and the image data is padded and data enhancement operations are performed to ensure that the length and width of the image are the same and not distorted.

[0021] Specifically, due to the complex environment in which the fire occurs and the changing surrounding lighting conditions, Mosaic data enhancement is performed on the image data at this time. The image data is spliced ​​by random scaling, random cropping, and random arrangement, which greatly enriches the background information.

[0022] The garage scene is monitored by video capture in real time, and the captured video is processed into frame images. The images are re-padding with gray bars and zero-filled to a size of 460×460: when the length of the image is less than 460, the length is padded with gray bars; when the width of the image is less than 460, the width is padded with gray bars and zero-filled to ensure that the length and width of the image are the same, so that the fire smoke and flame image is not distorted. Finally, the image processed with gray bars and zero-filled is sent to the fire recognition model.

[0023] Step 2: Input the preprocessed image data into the fire recognition model to obtain the fire recognition results and locate the fire point; Specifically, the preprocessed image data is input into the fire recognition model based on the YOLOv5s model. The fire recognition model consists of three parts: Backbone network, Neck network and Prediction. The Backbone network of the fire recognition model uses deep separable convolution, and the Backbone network is the backbone network; deep separable convolution can greatly reduce the number of parameters of the fire recognition model; in order to ensure the accuracy, real-time and deployment requirements of fire detection, the backbone network design does not use residual connections and other operations that increase model parameters, and a small number of attention modules are added to the bottom layer of the backbone network; an attention mechanism that focuses on the number of channels is added to the bottom layer of the Backbone network, and a parameter-free attention mechanism is introduced in the Neck network. The parameter-free attention mechanism consists of activation functions and energy functions connected in sequence.

[0024] First, the preprocessed image data is input into the fire recognition model based on the YOLOv5s model, and then feature extraction is performed through the Backbone network. After multiple convolution modules, a 64×64 feature map is obtained. The 64×64 feature map is passed through multiple convolution layers to obtain a 128×128 feature map. The 128×128 feature map is further convolved to obtain a 256×256 feature map. The image is processed through layer-by-layer convolution, point-by-point convolution, global average pooling layer, first fully connected layer and second fully connected layer. In this embodiment, only the first fully connected layer and the second fully connected layer are used as optional elements to balance the recognition efficiency and accuracy of the fire recognition model and better extract the characteristics of fire smoke and flame. At this time, the layer-by-layer convolution layer, the point-by-point convolution layer, and the first fully connected layer use the SiLU activation function, and the second fully connected layer uses the hsigmoid activation function. The optional option is the attention mechanism that focuses on the number of channels.

[0025] The 64×64 feature map is used to detect large fire targets, the 128×128 feature map is used to detect medium fire targets, and the 256×256 feature map is used to detect small fire targets; the three feature maps are then input into the Neck network for further enhanced feature extraction, and the fire smoke and flame images are enhanced with the parameter-free attention mechanism; the parameter-free attention mechanism consists of sequentially connected activation functions and energy functions; the SPPF fast feature pyramid includes a convolutional layer, a first maximum pooling layer, a second maximum pooling layer, a third maximum pooling layer, and a fusion layer connected in sequence using multiple residuals. The pooling kernels of the first maximum pooling layer, the second maximum pooling layer, and the third maximum pooling layer are 5×5, 9×9, and 13×13, respectively.

[0026] After feature fusion, the 256×256 feature map is convolved through the second convolution layer and upsampled through the first upsampling layer to obtain an upsampled feature map. The upsampled feature map is fused with the 128×128 feature map to obtain a first feature fusion map. The first feature fusion map is sequentially passed through the convolution layer and the upsampling layer and then fused with the 64×64 feature map to obtain a second feature fusion map. The second feature fusion map is sequentially convolved to obtain a 64×64 output feature map. The second feature fusion map is fused with the first feature fusion map to obtain a third feature fusion map. The third feature fusion map is subjected to a feature fusion operation with the 256×256 feature map after the second convolution layer to obtain a fourth feature fusion map; the fourth feature fusion map is subjected to convolution processing to obtain a 256×256 output feature map; The 64×64 output feature map, 128×128 output feature map, and 256×256 output feature map are used for target detection. If a fire occurs, non-maximum suppression based on the CIOUloss loss function with the introduction of a penalty coefficient is used, and the confidence threshold is set to 0.5, so as to obtain the prediction box with the highest confidence and realize accurate identification of the fire point.

[0027] Step 3: Obtain fire data of the fire starting point, input the fire data into the situation development prediction model, and obtain the fire development prediction result; based on the fire development prediction result, use the support vector machine algorithm to identify the potential fire risk and risk level in the scene, and generate the scene risk assessment result; Specifically, the fire data includes fire temperature, CO concentration, visibility, and smoke recirculation length; The key feature analysis of these four data factors is performed. The Lasso method is a compression estimation method based on the idea of ​​reducing the variable set. It constructs a penalty function to compress the coefficients of the variables and make some regression coefficients become 0, thereby achieving the purpose of variable selection. Using Lasso to screen variables can reduce the complexity of the model, solve the problem of data collinearity, and avoid overfitting. Therefore, Lasso regression is used to select key factors.

[0028] The specific internal meaning of Lasso regression is to add L1 regularization after the loss function to perform variable selection. L1 regularization is shown in the following formula:

[0029] in, m is the number of data factors, k is the number of parameters, i For a certain data factor, is L1 regularization, h is the coefficient.

[0030] The situation development prediction model is a generalized regression neural network GRNN. The traditional GRNN network structure has four layers, namely input layer, pattern layer, summation layer and output layer. The present invention adopts the firefly algorithm to optimize the GRNN network structure, combines the global search capability of the firefly algorithm and the regression characteristics of the generalized regression neural network, and optimizes the GRNN neural network so that the model performance reaches the optimal parameter combination.

[0031] Furthermore, the firefly algorithm is used to train and optimize the generalized regression neural network GRNN, and the optimization process of the optimal parameter combination is obtained as follows: 1) Parameter initialization, including the initial number of fireflies (i.e., search breadth), the number of iterations (i.e., search depth), the brightness of the fireflies (i.e., attractiveness), and the initial position and initial brightness of the fireflies in the search space.

[0032] 2) Data input: importing data sets for training and optimization into the algorithm; 3) Firefly position initialization: The initial position of the firefly in the search space is randomized. The position of the fireflies represents a possible combination of parameters for the generalized regression neural network model. This randomized initialization method helps the algorithm to search evenly in the search space.

[0033] 4) Attraction evaluation: The brightness of a firefly reflects the performance of the generalized regression neural network correction model under its corresponding parameter combination. The higher the brightness, the better the performance of the model under this parameter combination. Based on the brightness of a firefly and the brightness of other fireflies, the attraction between fireflies can be calculated. The size of the attraction determines the direction and distance of the firefly's movement in the search space.

[0034] 5) Position update: Based on the firefly's attractiveness and some random disturbance factors, the position of the firefly in the search space is updated. This position update is essentially a fine-tuning of the generalized regression neural network correction model parameter combination, aiming to find a better solution.

[0035] 6) Brightness update and evaluation: After the firefly position is updated, the brightness of the firefly needs to be re-evaluated. That is, calculate the performance of the updated generalized regression neural network correction model on the data set.

[0036] 7) Termination condition judgment: After each iteration, it is necessary to check whether the preset termination condition is met.

[0037] These termination conditions include reaching the maximum number of iterations, meeting certain accuracy requirements, or the model performance not significantly improving after multiple consecutive iterations. If the termination conditions are met, the algorithm proceeds to the next step; otherwise, it returns to step 6) to continue execution.

[0038] 8) Result output: When the termination condition is met, the algorithm will output the training results. These results include the optimal generalized regression neural network correction model parameter combination and the corresponding model performance.

[0039] The present disclosure optimizes the parameters of the generalized regression neural network through the global search capability of the firefly algorithm, thereby finding the parameter combination that optimizes the model performance. Throughout the process, the brightness of the fireflies represents the performance of the model under the corresponding parameter combination, and the higher the brightness, the better the performance. Through the attraction and position update mechanism of the fireflies, the algorithm can continuously search for better solutions in the search space until the termination condition is met.

[0040] Furthermore, the specific processing of the input layer, pattern layer, summation layer and output layer of the generalized regression neural network GRNN is as follows: (1) The input layer only sends sample variables to the hidden layer and does not participate in the actual calculation.

[0041] (2) The number of neurons in the pattern layer is equal to the number of learning samples n, and each neuron corresponds to a different sample. Its transfer function is:

[0042] In the formula, is the smoothing factor, X is the sample input set; X i For the i The learning samples corresponding to the factors are i =1,2,.. m .

[0043] (3) The summation layer is used to sum the neurons in the pattern layer; (4) The output layer outputs the model results, neurons j The corresponding estimation results are j The elements are y j , whose estimated value is,

[0044] By outputting the estimated value, the next step of fire data is obtained, thereby obtaining the fire development prediction result.

[0045] Step 4: Based on the scenario risk assessment results, a firefighting relationship diagram between firefighting equipment, vehicles and fire points is constructed, and the dynamic relationship between entities is analyzed through the graph attention network to generate a dynamic relationship diagram; Based on the dynamic relationship diagram, using data visualization technology, simulate the scene and obtain real-time data changes, build a digital twin model to update and dynamically display data in real time, and generate a visualization interface; Specifically, based on the fire development prediction results, continuous homology group analysis is used to analyze the fire data structure by constructing and analyzing the multi-scale topological summary of the data, including life cycle and cave characteristics, extracting the key topological features in the data set and generating a topological feature set.

[0046] Furthermore, through continuous homology group analysis, multi-scale topological summary construction and analysis of data are performed for the fire development prediction results, including the extraction of life cycle and cave characteristics of the fire data set, and the use of continuous homology group theory to identify key topological structures in the data, such as by calculating the Betti number in the multidimensional data set, capturing the number of caves and connectivity of the data, and how the characteristics change with scale. This process involves constructing a series of nested subspaces, called filters, calculating the homology groups of the subspaces, and then generating a topological feature set, which summarizes the important topological structure information in the data set; accurately extracting the topological features in the data set that are critical to fire prediction, and then providing accurate basic data for pattern recognition, ensuring the high relevance and accuracy of the analysis results.

[0047] A machine learning classification algorithm combining vector machines and decision trees is used to identify potential fire development risks in firefighting scenarios. The existing support vector machine (SVM) is used to process multidimensional topological features. By selecting appropriate kernel functions, such as radial basis function (RBF), the feature space is transformed to improve the classification resolution. The decision tree algorithm constructs classification rules based on the classification results of SVM, and optimizes branch conditions using standards such as information gain to ensure that the generated decision tree model can efficiently identify different types of potential fire development risks. Through a precise classification mechanism, the identification of potential fire development risks in firefighting scenarios is more accurate, providing a key basis for the subsequent risk assessment of fire situation.

[0048] Furthermore, the present disclosure first performs a preliminary identification of the fire, and after confirming it as a fire, predicts the development of the fire situation based on the fire data, and uses the existing support vector machine (SVM) and decision tree algorithm to predict the potential risks in the fire development process based on the prediction results, and then performs the next step of risk assessment based on the prediction, identification and classification results, as follows: A risk assessment model based on fire probability is adopted, combined with statistical analysis methods and Bayesian networks, to evaluate the risk level in firefighting scenarios. By constructing a Bayesian network model, the potential fire pattern recognition results are used as input variables, and various potential risk factors and conditional probabilities are comprehensively considered. Statistical analysis methods such as conditional probability tables and prior probabilities are used to estimate and calculate the probability of occurrence of multiple types of risk events in different scenarios. The model can dynamically adjust and optimize the risk assessment framework, incorporating the accuracy of statistics and Bayesian theory. Through this method, an accurate assessment of different risk levels in firefighting scenarios can be generated, providing a scientific and precise basis for firefighting decision-making.

[0049] Furthermore, based on the results of scenario risk assessment, a graph theory algorithm is used to construct an algorithm using adjacency matrix and edge list to encode and structure the relationship between firefighting facilities, personnel and fire sources, capture the dynamic relationship between entities in the firefighting scene, and generate a firefighting relationship graph. Based on the firefighting relationship graph, the existing graph attention network is used to analyze the dynamic relationship between entities in the firefighting scene. By introducing the attention mechanism, differentiated weights are dynamically assigned to the nodes in the graph to generate a key feature relationship graph. Based on the key feature relationship graph, the dynamic graph update technology is used to update the time decay model and event-triggered algorithm, and the nodes and edges in the graph are adjusted and updated in real time to reflect the real-time status of the firefighting scene and generate a dynamic relationship graph.

[0050] Specifically, through graph theory algorithms, the relationships between firefighting equipment, vehicles and fire sources are encoded and structured. The adjacency matrix and edge list construction algorithm are used to convert the interaction relationship between entities in the firefighting scene into a graph expression. The adjacency matrix is ​​a two-dimensional array used to represent whether there are edges between nodes, and the edge list lists all node pairs and their relationships. Through the two data structures, the dynamic relationships in the firefighting scene are accurately captured and encoded to generate a firefighting relationship graph. It not only realizes the visualization of the entity relationship in the firefighting scene, but also lays a structured foundation for subsequent in-depth analysis. The generated firefighting relationship graph describes in detail the interaction relationship between facilities, vehicles and fire sources, providing an intuitive and structured tool for understanding and analyzing firefighting scenes.

[0051] Based on the generated fire relationship graph, the Graph Attention Network (GAT) is used for in-depth analysis to study the dynamic relationship between entities in the fire scene. GAT introduces the attention mechanism to dynamically assign different weights to each node in the graph, thereby highlighting the key nodes and edges; GAT evaluates the importance of each node to its adjacent nodes, adjusts the representation of node features through the learned weight parameters, and then generates a key feature relationship graph; it can identify the entities and interactions that are most critical to fire scene analysis, and optimize the focus and efficiency of subsequent analysis; the generated key feature relationship graph provides an accurate basis for fire decision-making, highlights the areas and entities that need to be focused on, and enhances the pertinence and effectiveness of fire scene analysis.

[0052] Dynamic graph update technology is used to adjust and update the nodes and edges in the graph in real time to reflect the latest status of the fire scene. The time decay model and event trigger algorithm are used to dynamically adjust the structure in the graph. The time decay model automatically adjusts the weights of nodes and edges according to time changes to simulate the dynamic changes in the relationship between entities. The event trigger algorithm updates the graph structure in real time according to new events in the fire scene, such as adding new nodes or edges, or adjusting the properties of nodes and edges. The dynamic update mechanism ensures that the fire relationship graph can promptly reflect changes in the fire scene. The generated dynamic relationship graph provides a real-time updated decision support tool for fire monitoring and emergency response, thereby improving the response speed and accuracy of fire management.

[0053] Furthermore, data visualization technology is used to simulate firefighting scenes through three-dimensional graphics, and dynamic charts are used to reflect data changes in real time, track the dynamics of key indicators, and generate data visualization displays.

[0054] Adopt user interface interaction design principles, including responsive design and user-oriented interaction strategies, to optimize the user experience, obtain and analyze required information, and generate an interactive design interface.

[0055] Adopt interface performance optimization technology, reduce interface loading time through performance optimization, improve the fluency of user interaction, use visual effect enhancement method design and animation effects to enhance interface attractiveness, and generate a comprehensive visual interface.

[0056] Step 5: Deploy and control firefighting actions based on the visual interface to achieve precise fire extinguishing in the garage.

[0057] Specifically, advanced graphics rendering technologies such as WebGL can be used in combination with data visualization libraries such as D3.js and Three.js to construct three-dimensional firefighting scenes and dynamic charts; three-dimensional graphics simulate the spatial layout of firefighting resource allocation in detail, such as the location of firefighting vehicles and rescue personnel, fire source distribution, etc., while dynamic charts reflect changes in key indicators such as response time and fire control progress in real time; visual representation simplifies complex data information, helping decision makers and relevant personnel to quickly understand scene conditions and resource dynamics, and improve information acceptability and decision-making efficiency; the generated data visualization display provides users with an intuitive and interactive interface to promote more effective resource management and scheduling decisions.

[0058] Example 2 In one embodiment of the present disclosure, a fire control and display system based on digital twin is provided, including: A data acquisition module, used for acquiring image data in the garage and preprocessing the image data in the garage; The fire identification module is used to input the pre-processed image data into the fire identification model to obtain the fire identification result and locate the fire point; An assessment module is used to obtain fire data of the fire point, input the fire data into the situation development prediction model, and obtain the fire development prediction result; based on the fire development prediction result, a support vector machine algorithm is used to identify the potential fire risk and risk level in the scene, and generate a scene risk assessment result; The model building and visualization module is used to build a firefighting relationship diagram between firefighting equipment, vehicles and fire points based on the results of scenario risk assessment, analyze the dynamic relationship between entities through the graph attention network, and generate a dynamic relationship diagram; based on the dynamic relationship diagram, use data visualization technology to simulate the scenario and obtain real-time data changes, build a digital twin model for real-time update and dynamic display of data, and generate a visualization interface; The fire control module is used to deploy and control firefighting actions based on a visual interface to achieve precise fire extinguishing in the garage.

[0059] Example 3 In one embodiment of the present disclosure, a computer program product is provided, including a computer program, which, when executed by a processor, implements the fire point fire control display method based on digital twins.

[0060] Example 4 In one embodiment of the present disclosure, a non-transitory computer-readable storage medium is provided, wherein the non-transitory computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by a processor, the fire control display method of the fire point based on digital twin is implemented.

[0061] Example 5 In one embodiment of the present disclosure, an electronic device is provided, including: a processor, a memory, and a computer program; wherein the processor is connected to the memory, and the computer program is stored in the memory. When the electronic device is running, the processor executes the computer program stored in the memory, so that the electronic device executes the fire control display method based on digital twin.

[0062] The present disclosure is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present disclosure. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0063] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0064] Although the above describes the specific implementation methods of the present disclosure in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present disclosure. Technical personnel in the relevant field should understand that on the basis of the technical solution of the present disclosure, various modifications or variations that can be made by those skilled in the art without creative work are still within the scope of protection of the present disclosure.

Claims

1. A fire control display method based on digital twins, characterized in that: include: Acquire image data in the garage, and pre-process the image data in the garage; The pre-processed image data is input into the fire recognition model to obtain the fire recognition result and locate the fire point; Obtaining fire data of the fire starting point, inputting the fire data into a situation development prediction model, and obtaining a fire development prediction result; Based on the fire development prediction results, the support vector machine algorithm is used to identify the potential fire risks and risk levels in the scene and generate the scene risk assessment results; Based on the results of scenario risk assessment, a firefighting relationship diagram between firefighting equipment, vehicles and fire points is constructed, and the dynamic relationship between entities is analyzed through the graph attention network to generate a dynamic relationship diagram; Based on the dynamic relationship diagram, using data visualization technology, simulate the scene and obtain real-time data changes, build a digital twin model to update and dynamically display data in real time, and generate a visualization interface; Firefighting actions are deployed and controlled based on the visual interface to achieve precise fire extinguishing in the garage.

2. The fire control display method based on digital twin according to claim 1, characterized in that: By shooting a video inside the garage, the video is processed into frame image data, and the image data is padded and data augmented to ensure that the length and width of the image are the same and not distorted.

3. The fire control display method based on digital twin according to claim 1, characterized in that: The preprocessed image data is input into the fire recognition model based on the YOLOv5s model. The fire recognition model consists of three parts: Backbone network, Neck network and Prediction. The Backbone network of the fire recognition model uses deep separable convolution, and adds an attention mechanism that focuses on the number of channels at the bottom of the Backbone network. A parameter-free attention mechanism is introduced in the Neck network. The parameter-free attention mechanism consists of activation functions and energy functions connected in sequence.

4. The fire control display method based on digital twin according to claim 1, characterized in that: The fire data of the ignition point is obtained, and the fire data includes fire temperature, CO concentration, visibility and smoke recirculation length. Lasso regression is used to select key features, and the fire situation of the ignition point is predicted using the situation development prediction model GRNN to generate a fire development prediction result.

5. The fire control display method based on digital twin according to claim 1, characterized in that: Based on the fire development prediction results, continuous homology group analysis is used to analyze the fire data structure by constructing and analyzing the multi-scale topological features of the data, including life cycle and cave characteristics, and extracting data The key topological features are concentrated to generate a topological feature set; Based on the topological feature set, a machine learning classification algorithm is used, a vector machine is used to classify multi-dimensional features, and classification rules are constructed through a decision tree to identify potential fire risks in firefighting scenarios and generate potential fire pattern recognition results; Based on the results of potential fire pattern recognition, a risk assessment model is adopted to construct a risk assessment framework based on fire probability. Combined with statistical analysis methods and Bayesian networks, the probability of occurrence of multiple types of risk events is calculated and predicted, the risk level in fire scenarios is evaluated, and scenario risk assessment results are generated.

6. The fire control display method based on digital twin according to claim 1, characterized in that: Using graph theory algorithms, adjacency matrix and edge list construction algorithms, the relationships between firefighting equipment, vehicles and fire sources are encoded and structured, the dynamic relationships between entities in firefighting scenes are captured, and a firefighting relationship graph is generated; A graph attention network is used to analyze the dynamic relationships between entities in firefighting scenarios. By introducing the attention mechanism, differentiated weights are dynamically assigned to the nodes in the graph to generate a key feature relationship graph. Based on the key feature relationship graph, dynamic graph update technology is used to update the time decay model and event-triggered algorithm, and the nodes and edges in the graph are adjusted and updated in real time to reflect the real-time status of the firefighting scenario and generate a dynamic relationship graph.

7. The fire control and display system based on digital twin is characterized by: include: A data acquisition module, used for acquiring image data in the garage and preprocessing the image data in the garage; The fire identification module is used to input the pre-processed image data into the fire identification model to obtain the fire identification result and locate the fire point; An evaluation module is used to obtain fire data of the fire point, input the fire data into a situation development prediction model, and obtain a fire development prediction result; Based on the fire development prediction results, the support vector machine algorithm is used to identify the potential fire risks and risk levels in the scene and generate the scene risk assessment results; The model building and visualization module is used to build a firefighting relationship diagram between firefighting equipment, vehicles and fire points based on the scenario risk assessment results, analyze the dynamic relationship between entities through the graph attention network, and generate a dynamic relationship diagram; Based on the dynamic relationship diagram, using data visualization technology, simulate the scene and obtain real-time data changes, build a digital twin model to update and dynamically display data in real time, and generate a visualization interface; The fire control module is used to deploy and control firefighting actions based on a visual interface to achieve precise fire extinguishing in the garage.

8. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the fire control display method based on digital twin of the fire point according to any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium, characterized in that: The non-transitory computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by the processor, the fire control display method based on digital twin according to any one of claims 1 to 6 is implemented.

10. An electronic device, characterized in that: include: A processor, a memory and a computer program; wherein the processor is connected to the memory, and the computer program is stored in the memory. When the electronic device is running, the processor executes the computer program stored in the memory so that the electronic device executes the fire control display method based on digital twins as described in any one of claims 1 to 6.

Images