Fire data processing method and smart fire platform

Abstract

The present invention relates to the field of smart fire protection technology, and in particular to a fire protection data processing method and a smart fire protection platform. The method comprises the following steps: collecting fire supervision data; performing fire anomaly feature analysis based on the fire supervision data to generate fire anomaly feature data; performing simulated fire anomaly spatiotemporal evolution analysis of the fire supervision area based on the fire anomaly feature data to generate simulated fire anomaly spatiotemporal evolution data; performing spatiotemporal node evaluation processing of the fire risk index on the simulated fire anomaly spatiotemporal evolution data to generate a fire risk spatiotemporal evaluation index; performing multi-agent collaborative fire protection intelligent control parameter analysis based on the fire risk spatiotemporal evaluation index to generate multi-agent collaborative fire protection intelligent control parameters; and executing multi-agent collaborative fire protection intelligent control operations according to the collaborative multi-agent fire protection intelligent control parameters. The present invention enables multi-agent collaborative fire protection control to achieve precise fire protection measures by analyzing the evolution trend of fire.

Description

Technical Field

[0001] The present invention relates to the field of smart fire protection technology, and in particular to a fire protection data processing method and a smart fire protection platform. Background Art

[0002] The demand for fire data processing stems from the complexity of fire prevention and control work and the diversity of data. With the increasing complexity of modern buildings and urban environments, fire supervision data is large in volume and diverse in types, including video images, sensor data, and historical fire records. Smart fire protection utilizes advanced technologies such as the Internet of Things, big data, cloud computing, and artificial intelligence to build an integrated fire management system. By real-time monitoring, analysis, and processing of multi-source data collected by sensors, cameras, and other equipment, it achieves early warning, dynamic assessment, and intelligent control of fire risks. It automatically identifies fire hazards, predicts fire development trends, and optimizes emergency decision-making and resource allocation, thereby improving the efficiency and accuracy of fire safety management and providing comprehensive protection for urban safety. However, due to the diverse sources of fire supervision data, traditional fire data processing methods have difficulty in effectively integrating and analyzing these heterogeneous data. In addition, the extraction and analysis of fire anomaly characteristics are difficult to be accurate and real-time, and lack scientific quantitative model support. Therefore, it is impossible to achieve real-time dynamic fire decision optimization, resulting in fire safety risks in fire protection implementation. Summary of the Invention

[0003] Based on this, the present invention provides a fire data processing method and a smart fire platform to solve at least one of the above technical problems.

[0004] To achieve the above object, a fire data processing method includes the following steps:

[0005] Step S1: Obtain fire supervision area data; collect fire supervision data on the fire supervision area data to generate fire supervision data; perform fire supervision abnormality feature analysis based on the fire supervision data to generate fire supervision abnormality feature data;

[0006] Step S2: performing data decoupling processing on the fire supervision abnormality feature data to generate decoupled fire supervision abnormality feature data; performing fire abnormality feature extraction based on the decoupled fire supervision abnormality feature data to generate fire abnormality feature data;

[0007] Step S3: Perform digital spatial modeling of the fire supervision area based on the fire supervision area data to generate a fire supervision area model; transmit the fire anomaly characteristic data to the fire supervision area model to perform simulated fire anomaly spatiotemporal evolution analysis of the fire supervision area to generate simulated fire anomaly spatiotemporal evolution data; perform spatiotemporal node evaluation of the fire risk index on the simulated fire anomaly spatiotemporal evolution data to generate a fire risk spatiotemporal evaluation index;

[0008] Step S4: Obtaining a fire protection implementation rule decision; performing a fire protection implementation rule multi-criteria decision optimization analysis based on the fire protection implementation rule decision to generate optimized fire protection implementation rule multi-criteria decision data; establishing a fire protection implementation rule adaptive control engine based on the optimized fire protection implementation rule multi-criteria decision data;

[0009] Step S5: The fire risk spatiotemporal assessment index is transmitted to the fire implementation rule adaptive control engine to perform multi-agent collaborative fire intelligent control parameter analysis to generate multi-agent collaborative fire intelligent control parameters; and the multi-agent collaborative fire intelligent control operation is executed according to the collaborative multi-agent fire intelligent control parameters.

[0010] The present invention uses monitoring equipment and sensors to efficiently collect information within the fire monitoring area data, such as video, temperature, smoke concentration, etc. This step ensures that the fire protection system obtains key information in real time, thereby accelerating response speed and improving monitoring accuracy. Abnormal feature analysis based on fire monitoring data can identify potential fire risks at an early stage, such as abnormal temperature rise or rapid increase in smoke, thereby achieving preventive intervention and avoiding or reducing losses caused by fire. Data decoupling processing effectively separates overlapping data features, avoids interference between data, improves the accuracy and reliability of fire feature extraction, and extracts fire anomaly features from the decoupled fire monitoring abnormal feature data to more accurately locate the initial characteristics of the fire, providing a scientific basis for subsequent response measures. Digital spatial modeling provides a detailed digital model of the fire supervision area, simulating real-world environments and providing a simulation platform for fire prevention and response. Fire anomaly characteristic data is transferred to this model for simulated spatiotemporal evolution analysis. This generates detailed simulated spatiotemporal evolution data, enabling a better understanding of the dynamics of fire spread and optimizing the design of fire response measures. By performing spatiotemporal node-level fire risk index evaluation on this data, the resulting spatiotemporal fire risk assessment index provides a quantitative tool for assessing fire risk levels across different regions and time points, significantly enhancing the efficiency and responsiveness of the early warning system and the overall performance of the smart firefighting platform, making it more efficient and scientific in fire prevention and response. Multi-criteria decision-making optimization analysis of fire implementation rules, based on fire implementation rule decisions, enables the selection of the most appropriate firefighting measures based on different scenarios, optimizes resource allocation, and improves the efficiency and effectiveness of emergency response. An adaptive control engine for fire implementation rules, based on the optimized multi-criteria decision-making data, is established to dynamically adjust firefighting strategies based on real-time conditions, enhancing the system's adaptability and flexibility. The collaborative fire intelligent control parameter analysis of the multi-agent system can integrate data and resources from different agents to achieve efficient fire collaboration. Firefighting operations can be performed according to the fire intelligent control parameters optimized by the multi-agent system, which can quickly and effectively control the fire in actual fire situations and reduce casualties and property losses.

[0011] This specification provides a smart fire protection platform for executing the fire protection data processing method described above. The fire protection data processing method includes:

[0012] The intelligent fire supervision module is used to obtain fire supervision area data; collect fire supervision data from the fire supervision area data to generate fire supervision data; analyze fire supervision anomaly characteristics based on the fire supervision data to generate fire supervision anomaly characteristic data;

[0013] The fire anomaly analysis module is used to perform data decoupling processing on the fire supervision anomaly feature data to generate decoupled fire supervision anomaly feature data; extract fire anomaly features based on the decoupled fire supervision anomaly feature data to generate fire anomaly feature data;

[0014] The fire risk spatiotemporal assessment module is used to perform digital spatial modeling of the fire supervision area based on the fire supervision area data to generate a fire supervision area model; transmit the fire anomaly characteristic data to the fire supervision area model to perform simulated fire anomaly spatiotemporal evolution analysis in the fire supervision area to generate simulated fire anomaly spatiotemporal evolution data; and perform spatiotemporal node evaluation of the fire risk index on the simulated fire anomaly spatiotemporal evolution data to generate a fire risk spatiotemporal assessment index;

[0015] The fire protection implementation rule decision analysis module is used to obtain fire protection implementation rule decisions; perform fire protection implementation rule multi-criteria decision optimization analysis based on the fire protection implementation rule decisions to generate optimized fire protection implementation rule multi-criteria decision data; and establish a fire protection implementation rule adaptive control engine based on the optimized fire protection implementation rule multi-criteria decision data;

[0016] The multi-agent collaborative firefighting module is used to transmit the fire risk spatiotemporal assessment index to the fire implementation rule adaptive control engine to perform multi-agent collaborative firefighting intelligent control parameter analysis, generate multi-agent collaborative firefighting intelligent control parameters; and execute multi-agent collaborative firefighting intelligent control operations based on the collaborative multi-agent firefighting intelligent control parameters.

[0017] The beneficial effect of the present application is that the fire data processing method of the present invention can effectively solve the challenges of existing smart fire protection systems in data processing and analysis. Through multi-source data fusion technology, it can effectively integrate multi-source heterogeneous data from video images, sensor signals, historical fire data, etc., realize data standardization and consistency processing, and thus overcome the limitations of data heterogeneity integration. In addition, it adopts advanced feature extraction and data decoupling algorithms, which can accurately and in real time identify and analyze fire anomaly characteristics, significantly improve the efficiency and accuracy of data processing, and by establishing a scientific quantitative model for fire risk assessment and introducing an adaptive control engine, it can realize real-time dynamic optimization of fire protection decisions, reduce dependence on expert experience, thereby enhancing the intelligence level of the system, improving the effectiveness of fire warning and emergency response, and the implementation effect of fire protection strategies. Through the automated execution of the smart fire protection platform and the multi-agent collaborative strategy, it can better cope with complex fire situations, improve the collaborative combat capability and management level of the overall fire protection system, and ensure the fire prevention and control measures implemented by fire emergency. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A schematic flow chart of the steps of a fire data processing method according to the present invention;

[0019] Figure 2 for Figure 1 Detailed implementation steps of step S1 in FIG.

[0020] Figure 3 for Figure 1 Detailed implementation steps of step S3 in FIG.

[0021] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0022] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. It is obvious that the embodiments described are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts are within the scope of protection of the present invention.

[0023] In addition, the accompanying drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. Identical reference numerals in the figures denote identical or similar parts, and thus repetitive descriptions thereof will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor and / or microcontroller approaches.

[0024] It should be understood that although the terms "first," "second," and the like may be used herein to describe various elements, these elements should not be limited by these terms. These terms are used solely to distinguish one element from another. For example, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element, without departing from the scope of the exemplary embodiments. The term "and / or" as used herein includes any and all combinations of one or more of the listed associated items.

[0025] To achieve this, please refer to Figures 1 to 3 The present invention provides a fire data processing method, comprising the following steps:

[0026] Step S1: Obtain fire supervision area data; collect fire supervision data on the fire supervision area data to generate fire supervision data; perform fire supervision abnormality feature analysis based on the fire supervision data to generate fire supervision abnormality feature data;

[0027] Step S2: performing data decoupling processing on the fire supervision abnormality feature data to generate decoupled fire supervision abnormality feature data; performing fire abnormality feature extraction based on the decoupled fire supervision abnormality feature data to generate fire abnormality feature data;

[0028] Step S3: Perform digital spatial modeling of the fire supervision area based on the fire supervision area data to generate a fire supervision area model; transmit the fire anomaly characteristic data to the fire supervision area model to perform simulated fire anomaly spatiotemporal evolution analysis of the fire supervision area to generate simulated fire anomaly spatiotemporal evolution data; perform spatiotemporal node evaluation of the fire risk index on the simulated fire anomaly spatiotemporal evolution data to generate a fire risk spatiotemporal evaluation index;

[0029] Step S4: Obtaining a fire protection implementation rule decision; performing a fire protection implementation rule multi-criteria decision optimization analysis based on the fire protection implementation rule decision to generate optimized fire protection implementation rule multi-criteria decision data; establishing a fire protection implementation rule adaptive control engine based on the optimized fire protection implementation rule multi-criteria decision data;

[0030] Step S5: The fire risk spatiotemporal assessment index is transmitted to the fire implementation rule adaptive control engine to perform multi-agent collaborative fire intelligent control parameter analysis to generate multi-agent collaborative fire intelligent control parameters; and the multi-agent collaborative fire intelligent control operation is executed according to the collaborative multi-agent fire intelligent control parameters.

[0031] The present invention uses monitoring equipment and sensors to efficiently collect information within the fire monitoring area data, such as video, temperature, smoke concentration, etc. This step ensures that the fire protection system obtains key information in real time, thereby accelerating response speed and improving monitoring accuracy. Abnormal feature analysis based on fire monitoring data can identify potential fire risks at an early stage, such as abnormal temperature rise or rapid increase in smoke, thereby achieving preventive intervention and avoiding or reducing losses caused by fire. Data decoupling processing effectively separates overlapping data features, avoids interference between data, improves the accuracy and reliability of fire feature extraction, and extracts fire anomaly features from the decoupled fire monitoring abnormal feature data to more accurately locate the initial characteristics of the fire, providing a scientific basis for subsequent response measures. Digital spatial modeling provides a detailed digital model of the fire supervision area, simulating real-world environments and providing a simulation platform for fire prevention and response. Fire anomaly characteristic data is transferred to this model for simulated spatiotemporal evolution analysis. This generates detailed simulated spatiotemporal evolution data, enabling a better understanding of the dynamics of fire spread and optimizing the design of fire response measures. By performing spatiotemporal node-level fire risk index evaluation on this data, the resulting spatiotemporal fire risk assessment index provides a quantitative tool for assessing fire risk levels across different regions and time points, significantly enhancing the efficiency and responsiveness of the early warning system and the overall performance of the smart firefighting platform, making it more efficient and scientific in fire prevention and response. Multi-criteria decision-making optimization analysis of fire implementation rules, based on fire implementation rule decisions, enables the selection of the most appropriate firefighting measures based on different scenarios, optimizes resource allocation, and improves the efficiency and effectiveness of emergency response. An adaptive control engine for fire implementation rules, based on the optimized multi-criteria decision-making data, is established to dynamically adjust firefighting strategies based on real-time conditions, enhancing the system's adaptability and flexibility. The collaborative fire intelligent control parameter analysis of the multi-agent system can integrate data and resources from different agents to achieve efficient fire collaboration. Firefighting operations can be performed according to the fire intelligent control parameters optimized by the multi-agent system, which can quickly and effectively control the fire in actual fire situations and reduce casualties and property losses.

[0032] As an embodiment of the present invention, refer to Figure 1 The above is a flowchart of the steps of a fire data processing method of the present invention. In this embodiment, the fire data processing method includes the following steps:

[0033] Step S1: Obtain fire supervision area data; collect fire supervision data on the fire supervision area data to generate fire supervision data; perform fire supervision abnormality feature analysis based on the fire supervision data to generate fire supervision abnormality feature data;

[0034] In an embodiment of the present invention, a pre-set fire supervision area is obtained, including but not limited to a medical building area, a campus building area, etc., and environmental data of the fire supervision area is obtained in real time through surveillance cameras and various sensors (such as smoke detectors, temperature sensors, gas sensors, etc.) installed inside and outside the building, including video images, smoke concentration, temperature changes, combustible gas concentration, etc. The raw data obtained above is transmitted to the data acquisition system, and the data is cleaned, formatted and denoised by the data preprocessing module to generate structured fire supervision data. These data will be stored in the database for subsequent analysis, and the collected fire supervision data will be analyzed using machine learning algorithms (such as support vector machines, deep learning models, etc.) to identify abnormal features that are different from the normal state. For example, abnormal image features such as flames and smoke are detected by image processing technology; abnormal change trends in temperature or smoke concentration are detected by time series analysis to generate fire supervision abnormal feature data.

[0035] Step S2: performing data decoupling processing on the fire supervision abnormality feature data to generate decoupled fire supervision abnormality feature data; performing fire abnormality feature extraction based on the decoupled fire supervision abnormality feature data to generate fire abnormality feature data;

[0036] In an embodiment of the present invention, multivariate analysis technology (such as principal component analysis PCA or independent component analysis ICA) is used to perform data decoupling processing on the fire supervision abnormality feature data, converting complex multidimensional data into mutually independent single-dimensional data, and generating decoupled fire supervision abnormality feature data, aiming to reduce data redundancy, and perform fire abnormality feature analysis based on the characteristics of each dimensional data, thereby improving the accuracy of feature extraction. Fire abnormality feature extraction is performed on the decoupled fire supervision abnormality feature data. Cluster analysis algorithms (such as K-means, DBSCAN, etc.) are applied to identify abnormal clusters in the data, extract key features related to fire, such as high temperature areas, rapidly increasing smoke concentrations, or abnormal concentration changes of specific gases, and generate fire abnormality feature data, which will be used for further fire risk assessment and model analysis.

[0037] Step S3: Perform digital spatial modeling of the fire supervision area based on the fire supervision area data to generate a fire supervision area model; transmit the fire anomaly characteristic data to the fire supervision area model to perform simulated fire anomaly spatiotemporal evolution analysis of the fire supervision area to generate simulated fire anomaly spatiotemporal evolution data; perform spatiotemporal node evaluation of the fire risk index on the simulated fire anomaly spatiotemporal evolution data to generate a fire risk spatiotemporal evaluation index;

[0038] In an embodiment of the present invention, a three-dimensional digital model of the fire supervision area is constructed using digital modeling technology based on the building's design drawings, interior layout, and relevant data about the fire supervision area (e.g., the building's three-dimensional CAD drawings, laser scanning data, building material information, and ventilation system layout). The building's three-dimensional spatial data and structural information are collected through technologies such as laser scanning, drone photography, and BIM (Building Information Modeling). Simultaneously, information about various internal equipment, combustible material locations, passageways, doors, and windows is obtained. The collected data is preprocessed to ensure consistent formatting and remove noise. The preprocessed data is imported into digital modeling software (such as AutoCAD, Revit, or Blender) to create an accurate three-dimensional model of the fire supervision area. This model includes the building's physical structure (e.g., walls, floors, ceilings) and interior layout (e.g., room layout, door and window locations, equipment and combustible material locations, etc.), as well as the ventilation system and other important fire safety factors. Fire-related attribute data for each area (e.g., combustible material type and quantity, equipment location and type, etc.) is associated with the digital model to ensure that the model accurately reflects the actual fire supervision area. Ultimately, a complete 3D fire supervision area model is generated, which can be used for subsequent fire simulation and risk assessment. Fire anomaly characteristic data (such as high-temperature areas and smoke density fluctuations) is imported into the 3D fire supervision area model. This data serves as initial conditions for simulating the starting point and characteristics of a fire. In fire simulation software (such as FDS (Fire Dynamics Simulator) or CFAST (Consolidated Model of Fire and Smoke Transport)), simulation parameters are set, including fire type, combustible material properties, initial fire source location, and fire growth rate. The fire simulation analysis is then initiated, simulating the spread of fire within the building, including flame spread, heat conduction, smoke movement, and toxic gas diffusion. The simulation displays the temporal evolution of the fire, including temperature fields, smoke density fields, and fire expansion areas. Simulated spatiotemporal evolution data of fire anomalies is generated. The simulation results are then output and visualized, showcasing the spatiotemporal dynamics of fire spread through 3D animations or heat maps, helping to identify high-risk areas and critical time points.Based on the characteristics of the fire supervision area and the type of fire, an appropriate fire risk assessment model, such as the Fire Risk Index Method (FRIM) and a statistical analysis model, is selected. The simulated spatiotemporal evolution data of fire anomalies are input into the fire risk assessment model. Combined with various fire characteristics in the area (such as temperature, smoke concentration, oxygen level, and toxic gas concentration), the fire risk index for each spatial point and time point is calculated. The model outputs the fire risk index and generates a spatiotemporal distribution map of fire risk, identifying high-risk areas and time points, which is provided to fire decision makers for optimizing firefighting strategies and emergency response plans.

[0039] Step S4: Obtaining a fire protection implementation rule decision; performing a fire protection implementation rule multi-criteria decision optimization analysis based on the fire protection implementation rule decision to generate optimized fire protection implementation rule multi-criteria decision data; establishing a fire protection implementation rule adaptive control engine based on the optimized fire protection implementation rule multi-criteria decision data;

[0040] In an embodiment of the present invention, a smart fire protection platform retrieves fire protection implementation rule decisions based on pre-set fire protection strategies and emergency response guidelines. These decisions include action plans and response strategies to be taken in different fire scenarios. A fire protection implementation rule decision database is established to store emergency response rules for various fire scenarios. These rules are based on historical fire data, expert experience, and fire protection regulations. When the system detects an abnormal fire or high-risk fire situation, it extracts the corresponding fire protection implementation rule decisions from the decision database based on the current fire type, location, and fire spread. Multi-criteria decision optimization analysis is performed on the retrieved fire protection implementation rule decisions to address the diversity and uncertainty of complex fire scenarios. A multi-criteria decision analysis model (such as the AHP (Analytical Hierarchy Process), TOPSIS (Topology of Approximately Ideal Solutions), or fuzzy comprehensive evaluation method) is used to optimize the current fire protection implementation rule based on multiple decision criteria (such as personnel safety, fire control speed, and equipment protection). Weights are assigned to different decision criteria based on the model outputs. For example, in crowded situations, personnel safety is prioritized; in areas with critical equipment, equipment protection is prioritized. Through calculation and weight adjustment, optimized fire protection implementation rule multi-criteria decision data is generated. This optimization decision takes into account the balance between different criteria to ensure the best response measures in various possible fire scenarios, in order to build an adaptive control engine that can dynamically adjust the firefighting strategy to adapt to the different stages of fire development. Develop an adaptive control algorithm that can adjust the operating parameters of firefighting equipment (such as water spray volume, fan speed, etc.) based on real-time fire data and optimization decisions. The algorithm uses reinforcement learning or adaptive fuzzy control technology to automatically learn and adjust the control strategy, compile the adaptive control algorithm into a software module, and integrate it into the control system of the smart firefighting platform. This module receives fire supervision data and optimization decision data in real time, generates corresponding control signals, and automatically adjusts the operating status of the firefighting equipment.

[0041] Step S5: The fire risk spatiotemporal assessment index is transmitted to the fire implementation rule adaptive control engine to perform multi-agent collaborative fire intelligent control parameter analysis to generate multi-agent collaborative fire intelligent control parameters; and the multi-agent collaborative fire intelligent control operation is executed according to the collaborative multi-agent fire intelligent control parameters.

[0042] In an embodiment of the present invention, the spatiotemporal fire risk assessment index is continuously transmitted to the control engine to ensure that the engine dynamically adjusts the control strategy based on the latest risk assessment data. Data transmission is carried out through a distributed network, and each distributed network node corresponds to the spatial information of the fire to ensure the real-time and efficiency of firefighting. A multi-agent system model is constructed to simulate the collaborative work of multiple firefighting equipment (such as intelligent fire extinguishing equipment, intelligent water sprinkler systems, dry ice spraying equipment, etc.). Each agent has specific functions and behavioral rules and can respond independently according to control parameters. The control parameters of different agents are optimized and analyzed using collaborative control algorithms (such as distributed optimization algorithms and cooperative game theory models). The algorithm takes into account the interactions between agents, resource constraints and dynamic changes in fires to ensure the optimal collaborative control strategy. Based on the analysis results, multi-agent collaborative firefighting intelligent control parameters are generated to determine the specific operations of each agent (such as the amount of fire extinguishing agent released, the movement path, the operation sequence, etc.). The intelligent control parameters are converted into specific operation instructions and sent to each intelligent agent (such as fire-fighting robots, intelligent valve controllers, etc.) through wireless communication or wired connection. The execution status of each intelligent agent and the development of the fire are monitored in real time. The control parameters are adjusted through sensor feedback information to ensure dynamic optimization and efficient execution of collaborative control. The operational coordination of each intelligent agent is optimized through intelligent scheduling algorithms to avoid resource waste and operational conflicts, ensuring the efficient and stable operation of the entire fire protection system.

[0043] Preferably, step S1 includes the following steps:

[0044] Step S11: Obtaining fire supervision area data;

[0045] Step S12: collecting fire supervision image frames from the fire supervision area data through monitoring equipment to generate fire supervision image frames;

[0046] Step S13: collecting fire supervision signal frames in the fire supervision area through sensor equipment to generate fire supervision signal frames;

[0047] Step S14: performing fire supervision data synchronization integration processing on the fire supervision image data frame and the fire supervision signal frame to generate fire supervision data;

[0048] Step S15: performing fire supervision abnormality feature analysis based on the fire supervision data to generate fire supervision abnormality feature data.

[0049] This invention captures fire surveillance area data, ensuring comprehensive data from the area, providing a foundation for subsequent analysis. This comprehensive data collection significantly improves the accuracy and timeliness of fire detection, enabling the system to rapidly respond to potential fire threats. Fire surveillance image frames captured by monitoring equipment provide detailed visual information within the area, including environmental changes at different points in time. Visual recording is crucial for analyzing the environmental conditions and dynamic changes of fires. Fire surveillance signal frames captured by sensor equipment, such as temperature, smoke concentration, and gas composition, provide key fire monitoring parameters, helping to accurately assess fire risk and providing an important basis for early fire detection. Synchronous integration of image and signal data frames effectively eliminates data silos and improves data utilization efficiency. Synchronous data processing helps more accurately identify the characteristics and patterns of fire occurrence, enabling more precise fire prediction and response. Analysis of abnormal characteristics based on the integrated fire surveillance data allows for the timely detection of early signs of fire, such as abnormal temperature rise or rapid smoke growth. This can significantly reduce fire response time and allow for the preemptive deployment of necessary firefighting measures, effectively controlling the scale of fires and minimizing potential losses.

[0050] As an embodiment of the present invention, refer to Figure 2 As shown, Figure 1 Detailed implementation steps of step S1 are shown in the flowchart. In this embodiment, step S1 includes:

[0051] Step S11: Obtaining fire supervision area data;

[0052] In this embodiment of the present invention, various monitoring devices and sensors installed inside and outside a building are used to obtain environmental and status data for a fire supervision area. This data includes, but is not limited to, building layout information, firefighting equipment distribution, aisle locations, and door and window status, to obtain fire supervision area data.

[0053] Step S12: collecting fire supervision image frames from the fire supervision area data through monitoring equipment to generate fire supervision image frames;

[0054] In this embodiment of the present invention, high-definition cameras positioned at strategic locations continuously capture real-time image frames from the fire monitoring area. These frames are captured at a rate of at least 30 frames per second, ensuring image clarity and continuity, capable of capturing every detail before and after a fire. Infrared cameras and thermal imagers are used to collect temperature distribution data within the fire monitoring area and visualize data in low-light conditions. Thermal imaging data can reveal temperature anomalies behind walls and objects, while infrared images can penetrate smoke to identify hidden fire sources or hotspots, generating fire monitoring image frames.

[0055] Step S13: collecting fire supervision signal frames in the fire supervision area through sensor equipment to generate fire supervision signal frames;

[0056] In this embodiment of the present invention, fire monitoring signals are collected in real time by various sensors installed in the monitoring area (such as smoke sensors, temperature sensors, and gas sensors). These sensors detect environmental parameters at a high sampling rate (e.g., more than 10 times per second), generating high-precision real-time signal data. The analog signal data collected by the sensors is converted into digital signal frames containing timestamps, sensor locations, sensor types, and their measured values ​​(such as temperature, smoke concentration, and gas concentration). Signal frame formatting ensures the uniformity and compatibility of data from different sensor types, generating fire monitoring signal frames.

[0057] Step S14: performing fire supervision data synchronization integration processing on the fire supervision image data frame and the fire supervision signal frame to generate fire supervision data;

[0058] In an embodiment of the present invention, the collected fire monitoring image frames and signal frames are time-synchronized to ensure time consistency between different data sources. The Network Time Protocol (NTP) is used to time-calibrate all monitoring equipment and sensors so that each data frame has a unified time reference. The synchronized image frames and signal frames are synchronized and integrated to generate fire monitoring data. Data synchronization and integration can combine multiple sensor data and image data, and synchronize the time series and space accurately to improve the ability to perceive the environmental status. For example, by combining temperature data and smoke concentration data, the possible location and type of fire can be identified. After data synchronization and integration processing, fire monitoring data is generated, including multi-dimensional data such as image, temperature, smoke concentration, and gas concentration.

[0059] Step S15: performing fire supervision abnormality feature analysis based on the fire supervision data to generate fire supervision abnormality feature data.

[0060] In an embodiment of the present invention, a fire supervision anomaly detection model is constructed using machine learning and deep learning algorithms (such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs)) and baseline information of fire supervision data. The model is trained based on the baseline information of fire supervision data and is capable of identifying and classifying various abnormal features (such as image features of smoke appearance, signal features of abnormal temperature rise, etc.). Fire supervision data is input into the anomaly detection model for real-time analysis. The model automatically identifies abnormal images in image frames and abnormal signals in signal frames, and generates fire supervision anomaly feature data. These feature data identify the area where the fire occurred and its severity. The detected abnormal feature data is extracted and marked to generate an abnormal feature dataset containing information such as the type of anomaly, time of occurrence, location, and intensity.

[0061] Preferably, step S15 includes the following steps:

[0062] Step S151: Perform fire supervision baseline analysis based on fire supervision data to generate fire supervision baseline data;

[0063] Step S152: extracting fire supervision deviation baseline features from the fire supervision data based on the fire supervision baseline data to generate fire supervision deviation baseline feature data;

[0064] Step S153: performing data enhancement processing on the fire supervision deviation baseline feature data to generate enhanced deviation baseline feature data;

[0065] Step S154: performing fire supervision abnormality feature analysis based on the enhanced deviation baseline feature data to generate fire supervision abnormality feature data.

[0066] The present invention determines the baseline data by analyzing the fire supervision data, and establishes a reference standard for the environmental state under normal conditions, which helps to distinguish between normal and abnormal states, thereby ensuring the accuracy of fire detection and avoiding false alarms and missed reports. Based on the fire supervision baseline data, deviation analysis of the supervision data helps to identify data points that are significantly different from the normal pattern. This deviation baseline feature extraction enables the system to capture potential abnormal events more accurately and provide key clues for early warnings. By performing enhanced processing on the deviation baseline feature data (such as data expansion, noise filtering, etc.), the robustness of the data can be improved, thereby making the fire prediction model more robust, which helps to improve the performance of the system in various environments, especially when the data volume is small or the quality is not high. By using the deviation baseline feature data that has undergone data enhancement, the specific abnormal features can be further analyzed and determined, which not only identifies the fact that the abnormality has occurred, but also parses the specific signal frames and image frames of the abnormality, providing decision support for subsequent response measures.

[0067] In an embodiment of the present invention, historical fire supervision data and real-time fire supervision data are obtained from fire supervision data. These data include readings of multiple sensors, such as temperature, smoke concentration, gas concentration, infrared and thermal imaging images, etc., which are used to construct and analyze the normal baseline of fire supervision. Historical data is used to construct a baseline model of the fire supervision area. Statistical methods (such as mean, standard deviation) and machine learning methods (such as Gaussian mixture model (GMM), autoencoder) are used to establish a baseline model under normal conditions. The construction of the baseline model needs to cover normal state data under various environmental conditions. Based on the constructed baseline model, the real-time data is analyzed to generate fire supervision baseline data. This baseline data represents the parameter range of the supervision area under normal circumstances, and is used to identify abnormal situations that deviate from the normal state. Compare the real-time fire supervision data with the baseline data, and calculate the deviation of each monitoring indicator. For example, calculating the difference between a temperature reading and its baseline temperature, or the deviation between smoke concentration and its baseline concentration, uses statistical metrics (such as Z-scores and anomaly scores) and feature extraction algorithms (such as PCA (principal component analysis), ICA (independent component analysis), or LDA (linear discriminant analysis) to extract key features from the calculated deviation data, generating fire supervision deviation baseline feature data. This feature data describes how real-time data deviates from the normal baseline and helps identify potential anomalies. The generated deviation baseline feature data is stored in a data management system and associated with metadata such as timestamps and monitoring locations to support subsequent analysis and processing. Before data augmentation, the deviation baseline feature data is cleaned and preprocessed to remove noise, fill in missing data, and normalize the data format to ensure data consistency and integrity. Data augmentation techniques (such as noise injection, data interpolation, and data expansion) are used to enhance the deviation baseline feature data. By increasing the diversity and complexity of the data, the model's ability to detect anomalies is enhanced. Through the data augmentation process, enhanced deviation baseline feature datasets are generated. These enhanced datasets include more abnormal patterns and features, which can improve the generalization and robustness of the anomaly detection model. An enhanced deviation baseline feature dataset is used to train a fire supervision anomaly feature detection model. Machine learning algorithms such as deep learning models (such as LSTM (Long Short-Term Memory Network), CNN (Convolutional Neural Network)), random forests, or support vector machines (SVM) can be used to build anomaly detection models. By learning patterns in the data, the model can identify anomalies that deviate from the normal baseline. The deviation baseline features of real-time fire supervision data are input into the trained anomaly detection model for real-time analysis. The model identifies anomaly patterns based on the input features and generates fire supervision anomaly feature data. For example, if smoke concentration rises sharply within a short period of time and exceeds the baseline range, the model will flag it as a potential fire anomaly. Based on the detection results, fire supervision anomaly feature data is generated, indicating the specific anomaly type, time of occurrence, location, and severity.

[0068] Preferably, step S2 includes the following steps:

[0069] Step S21: performing data decoupling processing on the fire supervision abnormality feature data to generate decoupled fire supervision abnormality feature data;

[0070] Step S22: performing decoupled data clustering pattern analysis on the decoupled fire supervision abnormality feature data based on a preset clustering algorithm to generate decoupled data clustering pattern data;

[0071] Step S23: extracting fire anomaly features from the decoupled fire supervision anomaly feature data according to the decoupled data clustering pattern data to generate fire anomaly feature data.

[0072] The present invention performs data decoupling processing on the abnormal characteristic data of fire supervision, separates the overlapping or interdependent data features, thereby reducing the redundancy and interference between data features, clarifying the independent influence of each feature, and laying a solid foundation for subsequent more accurate data processing and analysis. The decoupled data is analyzed using a preset clustering algorithm (such as K-means, DBSCAN, etc.) to effectively identify various patterns or groups existing in the data. This clustering pattern analysis helps to discover statistically significant trends and patterns from a large amount of decoupled data. These trends and patterns may be directly related to specific types of fire risks or abnormal behaviors. Further fire anomaly feature extraction is performed on the decoupled fire supervision abnormal characteristic data based on the decoupled data clustering pattern data, which can highly accurately locate and identify early signs of possible fire occurrences. Through the extracted fire anomaly features, the fire protection system can quickly respond to potential fires and deploy rescue resources in advance, thereby significantly improving the efficiency and effectiveness of emergency response.

[0073] In an embodiment of the present invention, the abnormal characteristic data of fire supervision includes multidimensional data such as smoke concentration, temperature, gas composition, image features, etc. collected by various sensors and monitoring equipment. These data reflect abnormal conditions in different dimensions and have complex interdependencies and redundant information. Decoupling processing algorithms such as principal component analysis (PCA), independent component analysis (ICA) or factor analysis (FA) are used to decouple the abnormal characteristic data to help reduce the dimension of the data, remove redundant information, and highlight independent and important abnormal features in the data. For example, ICA can be used to decompose independent abnormal signal components, such as high temperature and high smoke concentration. Through decoupling processing, decoupled fire supervision abnormal characteristic data are generated, which retain the independent characteristic components between different anomalies. Select a clustering algorithm for decoupled data, such as K-means clustering, DBSCAN (density-based clustering), or hierarchical clustering. Before applying the clustering algorithm, standardize the decoupled fire monitoring anomaly feature data (e.g., using Z-score standardization). Apply the preset clustering algorithm to divide the decoupled data into several cluster patterns. These cluster patterns represent different types or stages of fire anomaly development. The clustering results reveal significant differences between areas with high temperature and high smoke concentrations and areas with low oxygen and high carbon monoxide. Save the cluster analysis results as decoupled data cluster pattern data, identifying the cluster to which each data point belongs and its characteristic center (e.g., average temperature and average smoke concentration for each cluster). These cluster pattern data help understand the distribution and classification of different fire anomaly features and facilitate further extraction of specific fire characteristics. Design a feature extraction strategy to identify and distinguish different fire anomaly features within the decoupled data cluster patterns. For example, for identified clusters, the center point feature values ​​of each cluster and their deviation from the baseline data are extracted to identify the severity or type of fire in that cluster (such as initial fire, high-heat fire, chemical combustion, etc.). Feature extraction algorithms (such as maximum entropy models, support vector machines (SVMs), and random forests) are used to extract representative fire anomaly features from the cluster pattern data. SVMs are used to perform secondary classification on the cluster data, further dividing it into finer anomaly types, or random forests are used to screen out the most important feature variables for fire identification. Based on the extracted features, fire anomaly feature data is generated. This dataset contains key features of different fire types (such as flame height, type of burning material, and diffusion rate), as well as information such as their location and time of occurrence. For example, the combination of high heat sources and rapidly spreading smoke in a specific area can be identified and marked as a high-risk fire area.

[0074] Preferably, step S3 includes the following steps:

[0075] Step S31: performing fire supervision area digital twin analysis on the fire supervision area data to generate fire supervision area digital twin data;

[0076] Step S32: using digital modeling technology to perform digital space modeling processing on the digital twin data of the fire supervision area to generate a fire supervision area model;

[0077] Step S33: transmitting the fire anomaly characteristic data to the fire supervision area model to perform simulated fire anomaly space mapping of the fire supervision area to generate simulated fire anomaly space mapping data;

[0078] Step S34: performing fire diffusion-spatial correlation analysis based on the simulated fire anomaly spatial mapping data to generate fire diffusion spatiotemporal correlation data;

[0079] Step S35: performing simulated fire anomaly spatiotemporal evolution analysis based on the fire diffusion spatiotemporal correlation data and the simulated fire anomaly spatial mapping data to generate simulated fire anomaly spatiotemporal evolution data;

[0080] Step S36: Based on the fire risk index method, the fire risk index spatiotemporal node evaluation processing is performed on the simulated fire abnormal spatiotemporal evolution data to generate a fire risk spatiotemporal evaluation index.

[0081] The present invention creates a real-time updated and highly accurate virtual copy (digital twin) of the fire supervision area by performing digital twin analysis on the fire supervision area data. This enables the system to simulate various real-world scenarios in a virtual environment without directly intervening in the actual environment, thereby safely testing and predicting various fire situations. The digital twin provides a visual and interactive way to analyze and understand complex environmental dynamics, enabling better risk assessment and resource planning for firefighting. Digital modeling technology is used to process the digital twin data of the fire supervision area to create a detailed digital model that reflects the specific spatial layout and structural characteristics of the supervision area. This not only improves the pertinence and effectiveness of fire response strategies, but also allows for accurate understanding and mastery of various possible fire paths and diffusion patterns before a fire occurs. The accurate digital model provides a practical and specific scenario for fire training, which helps to conduct more realistic fire simulations. Integrating fire anomaly feature data into the fire supervision area model and performing simulated fire anomaly spatial mapping can simulate the actual spread of fire in a specific environment, identify key fire source locations and spread pathways, and provide a scientific basis for developing specific firefighting strategies and evacuation plans. By simulating different fire spread scenarios, the effectiveness of existing firefighting facilities and strategies can be evaluated, allowing for timely adjustment and optimization of emergency response measures to ensure that losses are minimized in the event of a real fire. By analyzing simulated fire anomaly spatial mapping data, the correlation between fire spread and spatial structure is explored, revealing how fire interacts with building layout, material properties, and spatial configuration. High-risk fire areas and critical spread pathways can be accurately identified, enabling the development of targeted fire prevention measures and escape routes. This provides a critical basis for emergency fire planning and helps pre-determine the most effective response strategy before a fire occurs, including rapid response in the early stages of a fire and the selection of firefighting strategies. Further simulation analysis using the spatiotemporal correlation data of fire spread generates data on the evolution of fire anomalies across different temporal and spatial dimensions. Spatiotemporal evolution analysis provides in-depth insight into the fire development process, enabling real-time tracking of fire progress and prediction of future trends. This directly supports real-time decision-making, such as adjusting evacuation routes and determining the deployment location of firefighting resources, to adapt to the dynamic changes of fires and improve the timeliness and effectiveness of emergency response. By applying the fire risk index method to analyze the simulated spatiotemporal evolution data of fire anomalies, a comprehensive spatiotemporal fire risk assessment index is generated. This index quantifies the degree of fire risk in different regions and time points, providing management nodes and quantitative management information for fire management. The fire risk index can prioritize high-risk areas in emergencies and locate target fire sources, maximizing the effectiveness of fire management, rationally allocating limited firefighting resources, and maximizing resource utilization efficiency and firefighting effectiveness.

[0082] As an embodiment of the present invention, refer to Figure 3 As shown, Figure 1 Detailed implementation steps of step S3 are shown in the flowchart. In this embodiment, step S3 includes:

[0083] Step S31: performing fire supervision area digital twin analysis on the fire supervision area data to generate fire supervision area digital twin data;

[0084] In an embodiment of the present invention, relevant data of the fire supervision area is collected, including the three-dimensional structural data of the building (such as CAD drawings, BIM models), real-time sensor data (such as temperature, smoke concentration, gas concentration), surveillance camera image data, etc., and these multi-source heterogeneous data are integrated to provide a basis for digital twin analysis. A virtual model of the fire supervision area is constructed using digital twin technology. The digital twin model is a high-precision virtual copy that can dynamically reflect the physical structure and real-time environmental conditions of the building. Modeling software (such as Autodesk Revit, Blender, etc.) is used to combine the three-dimensional building structure data with real-time sensor data to generate digital twin data of the fire supervision area. In the virtual environment, the dynamic update of the digital twin model is driven by real-time sensor data, so that it can reflect the environmental changes and fire characteristics in the fire supervision area in real time, including the current temperature distribution, smoke diffusion, equipment status, etc., providing a basis for subsequent spatial modeling and fire simulation.

[0085] Step S32: using digital modeling technology to perform digital space modeling processing on the digital twin data of the fire supervision area to generate a fire supervision area model;

[0086] In an embodiment of the present invention, digital modeling technology (such as computer-aided design (CAD) and building information modeling (BIM)) is used to further process the digital twin data to create an accurate three-dimensional digital model that describes in detail the physical structure, spatial layout, passages, and door and window locations of the fire supervision area. The initially constructed three-dimensional model is refined by adding details such as wall materials, furniture layout, and the specific locations of fire-fighting equipment (such as fire extinguishers and sprinkler heads). The geometric accuracy and level of detail of the model are optimized so that it can accurately reflect the real building environment and provide an accurate foundation for subsequent simulations. Ultimately, a high-precision three-dimensional fire supervision area model is generated, which not only contains the geometric and topological information of the building, but is also dynamically updated with real-time data links to reflect the current monitoring status and environmental conditions.

[0087] Step S33: transmitting the fire anomaly characteristic data to the fire supervision area model to perform simulated fire anomaly space mapping of the fire supervision area to generate simulated fire anomaly space mapping data;

[0088] In this embodiment of the present invention, fire anomaly characteristic data (such as high-temperature areas and smoke distribution) is transferred to a three-dimensional model of the fire monitoring area. This data includes the fire's location, temperature gradient, and smoke density changes. Fire simulation software (such as Fire Dynamics Simulator (FDS) or PyroSim) is used to configure simulation parameters, define fire source attributes (such as fire source type, burning material, and initial fire source location) and the environmental conditions of the simulation scenario (such as wind speed and ventilation conditions). The simulation software is then run to perform spatial mapping analysis of the fire anomalies, simulating the fire's diffusion path, temperature distribution, and smoke diffusion under different temporal and spatial conditions. This generates simulated fire anomaly spatial mapping data, demonstrating the fire's evolution in three-dimensional space.

[0089] Step S34: performing fire diffusion-spatial correlation analysis based on the simulated fire anomaly spatial mapping data to generate fire diffusion spatiotemporal correlation data;

[0090] In an embodiment of the present invention, spatial correlation analysis methods (such as spatial autocorrelation analysis and spatiotemporal statistical analysis) are used to study the spatial patterns and correlations of fire spread. For example, the Moran's I index is applied to analyze the spatial clustering of fire spread and identify fire hotspots. Based on simulated fire anomaly spatial mapping data, the spatial patterns of fire spread are analyzed. For example, the spatial architectural feature information of fire spread is analyzed to identify the diffusion path, speed, and impact range. Spatiotemporal analysis tools (such as ArcGIS and QGIS) are used, combined with the time dimension, to study the spatiotemporal evolution of fire spread. The analysis results are output as spatiotemporal correlation data of fire spread, describing the temperature, smoke concentration, and their changing trends at each spatial location during the fire spread process, such as the diffusion behavior of fire in space.

[0091] Step S35: performing simulated fire anomaly spatiotemporal evolution analysis based on the fire diffusion spatiotemporal correlation data and the simulated fire anomaly spatial mapping data to generate simulated fire anomaly spatiotemporal evolution data;

[0092] In this embodiment of the present invention, a fire spatiotemporal evolution model is established based on simulated fire anomaly spatial mapping data and fire spread spatiotemporal correlation data. The model uses time series analysis methods (such as ARIMA models and LSTM neural networks) to simulate the temporal dynamic evolution of fires. The spatiotemporal evolution model is then run to simulate the spread and evolution of fires at different time points. The model considers the influence of factors such as building structure, ventilation conditions, and fire source characteristics to generate simulated fire anomaly spatiotemporal evolution data, including fire spread rate, heat flux distribution, and smoke diffusion paths, providing detailed dynamic data for fire risk assessment and decision support.

[0093] Step S36: Based on the fire risk index method, the fire risk index spatiotemporal node evaluation processing is performed on the simulated fire abnormal spatiotemporal evolution data to generate a fire risk spatiotemporal evaluation index.

[0094] In an embodiment of the present invention, a fire risk index method (such as the FIRIM fire risk index model) is used to evaluate simulated fire anomaly spatiotemporal evolution data. This method comprehensively considers multiple factors, including fire severity, spread rate, and impact range, to calculate a fire risk index for each spatiotemporal node. Based on the simulated fire anomaly spatiotemporal evolution data, a fire risk assessment is performed for each spatiotemporal node. The assessment includes spatial fire information, such as the fire risk level (low, medium, or high) for each room, floor, or area, and analyzes the temporal dynamics of risk changes. The assessment results are output as a spatiotemporal fire risk assessment index, which provides a quantitative risk metric for each spatiotemporal node, identifies high-risk areas and time periods, and provides data support for fire emergency response and decision-making.

[0095] Preferably, step S34 includes the following steps:

[0096] Step S341: Acquire historical fire spread spatial characteristic data;

[0097] Step S342: performing fire spread comprehensive evaluation index evaluation analysis of various spatial factors on the historical fire spread spatial characteristic data to generate historical fire spread comprehensive evaluation index data;

[0098] Step S343: establishing a temporal correlation mapping relationship between fire spread and spatial features based on a preset spatiotemporal graph neural network algorithm to obtain a fire spread-spatiotemporal feature correlation prediction model;

[0099] Step S344: transmitting the simulated fire anomaly spatial mapping data to the fire spread-spatial-temporal feature correlation prediction model to perform spatiotemporal correlation analysis of fire spread, and generating fire spread spatiotemporal correlation data.

[0100] The present invention provides an empirical basis for fire analysis by collecting and collating spatial characteristic data of historical fire diffusion. Historical data enables the model to take into account actual fire situations, increasing the reality comparison and historical depth of the analysis. Comprehensive evaluation index evaluation and analysis of historical fire diffusion spatial characteristic data helps to quantify fire risks and diffusion characteristics. Through a systematic evaluation method, various spatial factors (such as building materials, layout, combustible distribution, etc.) are integrated to provide a comprehensive evaluation index system, which helps to accurately evaluate and predict the potential risks of future fires. The spatiotemporal graph neural network algorithm is used to establish a temporal correlation mapping relationship between fire diffusion and spatial characteristics, effectively capturing the complex spatial and temporal dependencies in the fire diffusion process. The graph neural network is particularly suitable for processing graph structure data, enabling the model to accurately simulate and predict fire behavior in complex building environments. The simulated fire anomaly spatial mapping data is transmitted to the prediction model for spatiotemporal correlation analysis, which can analyze and predict the path and speed of fire diffusion in real time, predict its diffusion trend at the early stage of the fire, provide accurate real-time data support for the fire department, and help formulate more effective response measures.

[0101] In this embodiment of the present invention, historical fire data is collected from fire management systems, building safety monitoring systems, and historical fire reports. This data includes the fire's starting location, spread path, spread speed, combustion area, temperature changes, and smoke spread. The data is collected from multiple scenarios, such as fire incidents in different building types (e.g., high-rise buildings, warehouses, and industrial facilities). The collected historical fire data is cleaned and preprocessed. Incomplete or noisy data is removed to ensure data accuracy and completeness. Data from different data sources are formatted and integrated to form a unified dataset of historical fire spread spatial characteristics. Key spatial features, such as the location of the fire origin, spread direction, spread speed, and affected area, are extracted from the cleaned data. These features are then standardized to enable comparison and analysis on a consistent scale. A comprehensive fire spread evaluation index system is established, defining key indicators for assessing fire spread characteristics, such as spread speed, heat release rate, smoke concentration, type of burning material, and the flammability of the affected area. Each indicator reflects a different fire spread characteristic. A multifactor analysis method is used to evaluate the historical fire spread spatial characteristic data. Combining multiple evaluation indicators, the comprehensive impact of each spatial factor on fire spread is calculated. For example, the analytic hierarchy process (AHP) is used to assess the impact of the type of burning material and fire temperature characteristics on the speed and extent of fire spread. The analysis results are output as comprehensive evaluation indicators for historical fire spread, including comprehensive diffusion characteristic values ​​for each fire event. Graph neural network (GNN) algorithms for spatiotemporal data modeling, such as the spatiotemporal graph convolutional network (ST-GCN) and the spatiotemporal graph attention network (ST-GAT), are used to establish a temporal correlation mapping relationship between fire spread and spatial features. GNNs effectively capture the dependencies and temporal dynamics between spatial features, making them suitable for modeling spatiotemporal correlations in fire spread. Historical fire spread comprehensive evaluation indicator data and spatial feature data are constructed into a spatiotemporal graph structure. Nodes represent spatial locations (e.g., different areas of a building), while edges represent relationships between spatial locations (e.g., distance, ventilation conditions, etc.). Node and edge features include fire spread characteristics and spatial attributes. The GNN model is trained using historical data, and model parameters are adjusted through supervised or semi-supervised learning methods to enable the model to accurately predict the spatiotemporal correlations of fire spread. The model was validated using cross-validation and a test set to assess its accuracy and generalization capabilities. Ultimately, a fire spread-spatial-temporal correlation prediction model was generated. This model can predict the fire spread path and temporal dynamics within a building based on the input spatial features and initial fire information. Simulated fire anomaly spatial mapping data (including initial fire location, temperature distribution, and smoke diffusion) was input into the fire spread-spatial-temporal correlation prediction model. The model then reasoned with this input data to calculate the spatiotemporal correlation of fire spread.A spatiotemporal correlation prediction model analyzes the spread and evolution of fires at different spatial locations and time points. This model takes into account the structural characteristics of buildings, fire characteristics, and environmental factors to simulate the spread and development of fires within a building. The model outputs the analysis results and generates spatiotemporal correlation data on fire spread, describing the spread patterns and dynamic changes of fires in spatial and temporal dimensions, including fire spread speed, temperature changes, and smoke concentration distribution in each spatial region.

[0102] Preferably, step S342 includes the following steps:

[0103] Performing hierarchical feature analysis on historical fire spread spatial feature data to generate historical fire spread spatial hierarchical feature data;

[0104] Based on the historical fire diffusion spatial hierarchical characteristic data, the spatial correlation judgment matrix of fire diffusion is designed to obtain the spatial correlation judgment matrix;

[0105] The fire diffusion comprehensive evaluation index data of historical fire diffusion spatial characteristic data are evaluated and processed based on the spatial correlation judgment matrix to generate historical fire diffusion comprehensive evaluation index data.

[0106] The present invention helps to systematically identify and classify key features in fire data, including fire source location, diffusion speed, affected area, etc., by performing hierarchical feature analysis on historical fire diffusion spatial feature data. This structured analysis method not only clearly demonstrates the hierarchical relationship between data, but also improves the data processing efficiency and accuracy of subsequent processing steps. Hierarchical processing clearly demonstrates the weight and importance of the impact of data features at each level on fire diffusion, making the logic of fire analysis more transparent and traceable. Based on the spatial hierarchical feature data of historical fire diffusion, a spatial correlation judgment matrix is ​​designed to effectively quantify the relationship and intensity of interaction between various spatial factors (such as building structure, material storage, ventilation system, etc.). This judgment matrix is ​​auxiliary information for analyzing fire behavior and formulating preventive measures. It provides a systematic method to evaluate the impact of various factors on fire diffusion. Accurate spatial correlation analysis helps to build a more accurate fire prediction model. By understanding how different factors interact with each other, the behavior of fire in similar environments can be better predicted. A spatial correlation judgment matrix is ​​used to evaluate the spatial characteristics of historical fire spread using comprehensive evaluation indicators, generating a comprehensive index system with multi-dimensional evaluation. This comprehensive evaluation index considers multiple factors and more comprehensively reflects the level of fire risk, providing a scientific basis for formulating prevention and control strategies and emergency response. Through this detailed comprehensive evaluation index, resource allocation and response strategies can be optimized based on the risk level of each area, ensuring that sufficient fire prevention and disaster relief resources are deployed in key areas, thereby minimizing the damage caused by fire.

[0107] In this embodiment of the present invention, spatial characteristic data on the spread of historical fire events is collected from a database, including building floor plans, fire starting points, fire spread paths, temperature changes in the burning area, and smoke diffusion patterns. A hierarchical analysis method (AHP) is then used to perform a hierarchical feature analysis on this historical fire spread spatial characteristic data. This method identifies and classifies spatial characteristics at different levels of the fire spread process, such as building structure, internal passageway distribution, and the location of combustible materials. During the AHP process, the fire spread characteristics are decomposed into multiple levels, each representing characteristics at different scales. For example, the target level (which provides a comprehensive assessment of the risk of building fire spread); the criterion level (which includes factors such as the likelihood of fire occurrence, the spatial characteristics of fire spread, and the effectiveness of firefighting equipment); and the indicator level (which includes specific spatial factors under each criterion, such as the complexity of room layout, combustible material density, passageway ventilation conditions, and fire isolation measures) are used. The results of the AHP analysis are then organized into historical fire spread spatial hierarchical characteristic data. This data details the key spatial characteristics of each level of the fire spread process. Based on hierarchical feature data, a spatial correlation judgment matrix was constructed. Spatial correlation factors were introduced into the judgment matrix and quantified using data-based analytical methods (such as spatial statistical analysis and machine learning models). Using historical fire data and spatial analysis tools (such as GIS), combined with machine learning algorithms (such as support vector machines and random forests), the impact of different spatial factors on fire spread was quantified, resulting in a spatial correlation judgment matrix. Multivariate statistical analysis methods (such as multiple regression analysis) and machine learning techniques (such as support vector machines) were applied to the spatial correlation judgment matrix to evaluate spatial factors such as the fire resistance rating of building materials and the degree of openness of spatial layout. The contribution of each spatial factor to fire spread characteristics was calculated. Using the AHP method, pairwise comparisons of each factor were performed to determine its relative importance and calculate a comprehensive score. The evaluation results were output as comprehensive evaluation indicators for historical fire spread, detailing the contribution and mutual influence of each spatial factor on fire spread characteristics during different historical fire events, providing data support for predicting future fire spread paths and optimizing emergency response strategies.

[0108] Preferably, step S4 includes the following steps:

[0109] Step S41: Obtain fire protection implementation rule decision;

[0110] Step S42: performing a fire protection implementation rule multi-criteria decision analysis on the fire protection implementation rule decision to generate fire protection implementation rule multi-criteria decision data;

[0111] Step S43: Optimizing the fire protection implementation rule multi-criteria decision data based on a preset reinforcement learning algorithm to generate optimized fire protection implementation rule multi-criteria decision data;

[0112] Step S44: Establishing a fire protection implementation rule adaptive control engine based on the optimized fire protection implementation rule multi-criteria decision data.

[0113] The present invention provides a basic framework and reference point for execution by obtaining current fire implementation rules and decisions, ensuring that all subsequent optimization and automation measures are based on actual and effective fire prevention strategies, thereby enhancing the practicality and compliance of the entire system. Multi-criteria decision analysis is performed on fire implementation rules, taking into account the multi-dimensional influencing factors of various fire prevention rules (such as safety, cost-effectiveness, response time, etc.), helping to formulate more comprehensive and balanced fire prevention strategies, and helping to identify the most effective combination of rules to ensure a rapid and effective response in emergency situations. By applying a preset reinforcement learning algorithm to optimize the multi-criteria decision data, the system can learn from historical data and simulation results and automatically adjust decision parameters to improve the accuracy and effectiveness of decisions. The reinforcement learning algorithm can find the optimal or near-optimal fire prevention strategy through continuous iteration, thereby improving the intelligence level of the system. Based on the optimized multi-criteria decision data of fire implementation rules, an adaptive control engine for fire implementation rules is established. This engine can adapt to different fire situations and environmental changes in real time and automatically adjust fire prevention measures. The application of the adaptive control engine greatly improves the flexibility and effectiveness of fire response, ensuring that the best response state is always maintained in a changing environment.

[0114] In an embodiment of the present invention, current fire protection implementation rule decisions are collected from the fire management system and emergency plan database. These decision data include strategies for responding to different types of fires, such as fire extinguisher usage rules, personnel evacuation routes, fire equipment operation plans, and area blockade strategies. This decision is established based on existing fire-related information. A rule-based decision framework is constructed to categorize and store response measures for different types of fires for easy retrieval and application. The framework includes decision conditions (such as fire type and fire scale) and response measures (such as fire extinguishing, evacuation, and alarm). Based on the different needs of fire response, a multi-criteria decision analysis system is established. The decision criteria include personnel safety, fire control speed, equipment protection, and resource consumption. Each criterion represents a different priority and response strategy. A multi-criteria decision analysis model based on TOPSIS is used to quantitatively analyze the fire protection implementation rules. Each decision criterion is assigned a weight to reflect its relative importance. By calculating the decision weights and scores, a comprehensive decision score is generated. Based on the analysis results of the multi-criteria decision model, multi-criteria decision data for fire protection implementation rules is generated. This data includes the comprehensive score and ranking of each decision option, which is used to guide specific fire emergency response strategies. A reinforcement learning environment based on firefighting scenarios is constructed, including a state space (e.g., fire location, scale, and occupant distribution), an action space (e.g., evacuation, fire extinguishing, and alarming), and a reward mechanism (e.g., maximizing personal safety, minimizing property damage, and optimizing response time). Algorithms such as Deep Q-Learning, Policy Gradient, or Proximal Policy Optimization (PPO) are used to simulate different fire scenarios to learn optimal firefighting implementation rules. The environment then outputs multi-criteria decision data based on the firefighting implementation rules optimized by the reinforcement learning algorithm. An adaptive control engine architecture is designed that receives optimized firefighting implementation rule data in real time and dynamically adjusts the control strategy based on real-time fire conditions. The engine comprises a decision module, a control module, and a feedback module, responsible for making decisions, executing actions, and obtaining feedback, respectively. The engine is based on the optimized decision data and a fuzzy control algorithm. The fuzzy control algorithm quantifies the fire severity and designs specific firefighting control parameters. The established engine is used to adjust the operating parameters of firefighting equipment (e.g., sprinklers, smoke exhaust systems, and fire extinguishers) in real time to implement optimal emergency response measures. At the same time, the engine uses sensors and monitoring equipment to monitor the development of the fire and the effectiveness of emergency measures in real time, and adjusts strategies in a timely manner to ensure the efficiency and effectiveness of emergency response.

[0115] Preferably, step S5 includes the following steps:

[0116] Step S51: transmitting the fire risk spatiotemporal assessment index to the fire implementation rule adaptive control engine for multi-agent fire intelligent control parameter analysis to generate multi-agent fire intelligent control parameters;

[0117] Step S52: Perform collaborative decision optimization analysis on the multi-agent fire protection intelligent control parameters based on collaborative learning technology to generate multi-agent collaborative fire protection intelligent control parameters;

[0118] Step S53: Execute the multi-agent collaborative fire intelligent control operation according to the collaborative multi-agent fire intelligent control parameters.

[0119] The present invention transmits the spatiotemporal fire risk assessment index to the fire implementation rule adaptive control engine, and performs multi-agent fire intelligent control parameter analysis, so that the operating parameters of each agent can be adjusted in real time according to the actual fire scene situation, ensuring the optimal configuration and scheduling of fire resources, and improving the response speed and effectiveness of firefighting operations. Through precise control parameter analysis, the firefighting team can deploy appropriate resources and personnel more quickly to deal with specific challenges at the fire scene in a targeted manner. The use of collaborative learning technology to perform decision-making optimization analysis on multi-agent fire intelligent control parameters promotes information sharing and strategy coordination among different agents, which not only improves the efficiency of collaboration among various agents, but also optimizes the overall firefighting strategy, ensuring the coordination and effectiveness of various measures in complex fire environments. Collaborative learning technology enables multi-agent systems to learn each other's operating strategies and effects, continuously adjust and improve, and thus respond more accurately to various unforeseen changes and challenges. Firefighting operations are performed according to the optimized collaborative multi-agent firefighting intelligent control parameters, ensuring that each agent plays its most effective role in the fire scene, thereby improving the efficiency and results of the overall firefighting operations. The collaborative operations between agents reduce the increase in fire hazards in fire management and improve safety. During the execution process, the agents can make instant adjustments based on real-time feedback and dynamic fire scene conditions.

[0120] In an embodiment of the present invention, a spatiotemporal fire risk assessment index (such as fire spread rate, temperature distribution, smoke concentration change, etc.) is transmitted to an adaptive control engine for fire implementation rules. The assessment index provides risk assessment data for the current fire situation in different temporal and spatial dimensions. In the adaptive control engine, multiple intelligent agent models (such as fire extinguishing robots, water sprinkler systems, smoke exhaust systems, emergency evacuation guidance systems, etc.) are constructed. Each intelligent agent has specific control objectives and operation strategies. For example, the fire extinguishing robot prioritizes fire source control, while the water sprinkler system automatically adjusts the water spray volume according to the smoke concentration. Based on the spatiotemporal fire risk assessment index, a multi-agent fire intelligent control parameter analysis is performed. A decision analysis algorithm (such as Markov decision process (MDP), dynamic programming (DP), etc.) is used to calculate the optimal control parameters of each intelligent agent under the current fire scenario. The analysis results are output as multi-agent fire intelligent control parameters. These parameters define the specific operation instructions and priorities of each intelligent agent under the current spatiotemporal conditions, providing a basis for subsequent intelligent agent collaborative decision optimization. A collaborative learning framework is constructed to enable multiple intelligent agents to share information and coordinate actions during fire response. The collaborative learning framework, comprising a state sharing module, a policy coordination module, and a decision optimization module, supports real-time communication and joint decision-making among multiple agents. Collaborative learning algorithms (such as cooperative multi-agent reinforcement learning and the joint policy gradient algorithm) are employed to collaboratively optimize the parameters of multi-agent firefighting intelligent control. The collaborative learning algorithm simulates various fire scenarios to learn the optimal collaborative strategy among the agents. During the collaborative learning process, each agent's operational strategy is optimized to achieve an overall optimal collaborative operation. The algorithm considers the interactions among agents and resource constraints (such as water availability and energy consumption), continuously adjusting control parameters to generate the optimal collaborative control strategy. The optimized multi-agent collaborative firefighting intelligent control parameters are then output. These parameters define the agent behaviors and operational priorities during the collaborative firefighting operation, ensuring overall coordination and efficiency of the fire response. The optimized multi-agent collaborative firefighting intelligent control parameters are converted into specific operational instructions and issued in real time to each agent (such as the fire-fighting robot, water spray system, and emergency evacuation guidance system). These instructions include movement path, operating speed, water spray intensity, evacuation route guidance, and more. Each agent performs collaborative operations based on the instructions received. The fire monitoring system and agent sensors monitor the operation of each agent and the development of the fire in real time. Based on the actual results and fire changes, the agent's control parameters and operational strategies are dynamically adjusted to ensure continuous and efficient firefighting operations.For example, if a fire suddenly spreads to a new area, the system will automatically adjust the coverage of the sprinkler system. At the same time, the fire-fighting robot will change its route to give priority to the newly emerged fire source. During the execution of collaborative operations, the system will continuously collect operational feedback and fire response effects from each intelligent agent, update the control strategy database, and further optimize the multi-agent collaborative operation model. Through continuous learning and optimization, the response capability and adaptability of the entire intelligent fire protection system will be improved.

[0121] This specification provides a smart fire protection platform for executing the fire protection data processing method described above. The fire protection data processing method includes:

[0122] The intelligent fire supervision module is used to obtain fire supervision area data; collect fire supervision data from the fire supervision area data to generate fire supervision data; analyze fire supervision anomaly characteristics based on the fire supervision data to generate fire supervision anomaly characteristic data;

[0123] The fire anomaly analysis module is used to perform data decoupling processing on the fire supervision anomaly feature data to generate decoupled fire supervision anomaly feature data; extract fire anomaly features based on the decoupled fire supervision anomaly feature data to generate fire anomaly feature data;

[0124] The fire risk spatiotemporal assessment module is used to perform digital spatial modeling of the fire supervision area based on the fire supervision area data to generate a fire supervision area model; transmit the fire anomaly characteristic data to the fire supervision area model to perform simulated fire anomaly spatiotemporal evolution analysis in the fire supervision area to generate simulated fire anomaly spatiotemporal evolution data; and perform spatiotemporal node evaluation of the fire risk index on the simulated fire anomaly spatiotemporal evolution data to generate a fire risk spatiotemporal assessment index;

[0125] The fire protection implementation rule decision analysis module is used to obtain fire protection implementation rule decisions; perform fire protection implementation rule multi-criteria decision optimization analysis based on the fire protection implementation rule decisions to generate optimized fire protection implementation rule multi-criteria decision data; and establish a fire protection implementation rule adaptive control engine based on the optimized fire protection implementation rule multi-criteria decision data;

[0126] The multi-agent collaborative firefighting module is used to transmit the fire risk spatiotemporal assessment index to the fire implementation rule adaptive control engine to perform multi-agent collaborative firefighting intelligent control parameter analysis, generate multi-agent collaborative firefighting intelligent control parameters; and execute multi-agent collaborative firefighting intelligent control operations based on the collaborative multi-agent firefighting intelligent control parameters.

[0127] The beneficial effect of the present application is that the fire data processing method of the present invention can effectively solve the challenges of existing smart fire protection systems in data processing and analysis. Through multi-source data fusion technology, it can effectively integrate multi-source heterogeneous data from video images, sensor signals, historical fire data, etc., realize data standardization and consistency processing, and thus overcome the limitations of data heterogeneity integration. In addition, it adopts advanced feature extraction and data decoupling algorithms, which can accurately and in real time identify and analyze fire anomaly characteristics, significantly improve the efficiency and accuracy of data processing, and by establishing a scientific quantitative model for fire risk assessment and introducing an adaptive control engine, it can realize real-time dynamic optimization of fire protection decisions, reduce dependence on expert experience, thereby enhancing the intelligence level of the system, improving the effectiveness of fire warning and emergency response, and the implementation effect of fire protection strategies. Through the automated execution of the smart fire protection platform and the multi-agent collaborative strategy, it can better cope with complex fire situations, improve the collaborative combat capability and management level of the overall fire protection system, and ensure the fire prevention and control measures implemented by fire emergency.

[0128] The present invention is therefore intended to be illustrative and non-restrictive in all respects, with the scope of the invention being defined by the appended claims rather than the foregoing description, and all changes that come within the meaning and range of equivalents of the application documents are intended to be embraced therein.

[0129] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. A fire data processing method, characterized in that: The following steps are involved: Step S1: Obtain fire supervision area data; collect fire supervision data on the fire supervision area data to generate fire supervision data; perform fire supervision abnormality feature analysis based on the fire supervision data to generate fire supervision abnormality feature data; Step S2: performing data decoupling processing on the fire supervision abnormality feature data to generate decoupled fire supervision abnormality feature data; Extract fire anomaly features based on the decoupled fire supervision anomaly feature data to generate fire anomaly feature data; Step S3 includes: Step S31: performing fire supervision area digital twin analysis on the fire supervision area data to generate fire supervision area digital twin data; Step S32: using digital modeling technology to perform digital space modeling processing on the digital twin data of the fire supervision area to generate a fire supervision area model; Step S33: transmitting the fire anomaly characteristic data to the fire supervision area model to perform simulated fire anomaly space mapping of the fire supervision area to generate simulated fire anomaly space mapping data; Step S34: performing fire diffusion-spatial correlation analysis based on the simulated fire anomaly spatial mapping data to generate fire diffusion spatiotemporal correlation data; wherein step S34 includes: Step S341: Acquire historical fire spread spatial characteristic data; Step S342: performing fire spread comprehensive evaluation index evaluation analysis of various spatial factors on the historical fire spread spatial characteristic data to generate historical fire spread comprehensive evaluation index data; Step S343: establishing a temporal correlation mapping relationship between fire spread and spatial features based on a preset spatiotemporal graph neural network algorithm to obtain a fire spread-spatiotemporal feature correlation prediction model; Step S344: transmitting the simulated fire anomaly spatial mapping data to the fire spread-spatial-temporal feature correlation prediction model to perform spatiotemporal correlation analysis of fire spread, thereby generating fire spread spatiotemporal correlation data; Step S35: performing simulated fire anomaly spatiotemporal evolution analysis based on the fire diffusion spatiotemporal correlation data and the simulated fire anomaly spatial mapping data to generate simulated fire anomaly spatiotemporal evolution data; Step S36: performing spatiotemporal node evaluation processing of the fire risk index on the simulated fire anomaly spatiotemporal evolution data based on the fire risk index method to generate a spatiotemporal fire risk evaluation index; Step S4: Obtaining a fire protection implementation rule decision; performing a fire protection implementation rule multi-criteria decision optimization analysis based on the fire protection implementation rule decision to generate optimized fire protection implementation rule multi-criteria decision data; establishing a fire protection implementation rule adaptive control engine based on the optimized fire protection implementation rule multi-criteria decision data; Step S5: The fire risk spatiotemporal assessment index is transmitted to the fire implementation rule adaptive control engine to perform multi-agent collaborative fire intelligent control parameter analysis to generate multi-agent collaborative fire intelligent control parameters; and the multi-agent collaborative fire intelligent control operation is executed according to the collaborative multi-agent fire intelligent control parameters.

2. The fire data processing method according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: Obtaining fire supervision area data; Step S12: collecting fire supervision image frames from the fire supervision area data through monitoring equipment to generate fire supervision image frames; Step S13: collecting fire supervision signal frames in the fire supervision area through sensor equipment to generate fire supervision signal frames; Step S14: performing fire supervision data synchronization integration processing on the fire supervision image data frame and the fire supervision signal frame to generate fire supervision data; Step S15: performing fire supervision abnormality feature analysis based on the fire supervision data to generate fire supervision abnormality feature data.

3. The fire data processing method according to claim 2, characterized in that: Step S15 includes the following steps: Step S151: Perform fire supervision baseline analysis based on fire supervision data to generate fire supervision baseline data; Step S152: extracting fire supervision deviation baseline features from the fire supervision data based on the fire supervision baseline data to generate fire supervision deviation baseline feature data; Step S153: performing data enhancement processing on the fire supervision deviation baseline feature data to generate enhanced deviation baseline feature data; Step S154: performing fire supervision abnormality feature analysis based on the enhanced deviation baseline feature data to generate fire supervision abnormality feature data.

4. The fire data processing method according to claim 1, characterized in that: Step S2 includes the following steps: Step S21: performing data decoupling processing on the fire supervision abnormality feature data to generate decoupled fire supervision abnormality feature data; Step S22: performing decoupled data clustering pattern analysis on the decoupled fire supervision abnormality feature data based on a preset clustering algorithm to generate decoupled data clustering pattern data; Step S23: extracting fire anomaly features from the decoupled fire supervision anomaly feature data according to the decoupled data clustering pattern data to generate fire anomaly feature data.

5. The fire data processing method according to claim 1, characterized in that: Step S342 includes the following steps: Performing hierarchical feature analysis on historical fire spread spatial feature data to generate historical fire spread spatial hierarchical feature data; Based on the historical fire diffusion spatial hierarchical characteristic data, the spatial correlation judgment matrix of fire diffusion is designed to obtain the spatial correlation judgment matrix; The fire diffusion comprehensive evaluation index data of historical fire diffusion spatial characteristic data are evaluated and processed based on the spatial correlation judgment matrix to generate historical fire diffusion comprehensive evaluation index data.

6. The fire data processing method according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: Obtain fire protection implementation rule decision; Step S42: performing a fire protection implementation rule multi-criteria decision analysis on the fire protection implementation rule decision to generate fire protection implementation rule multi-criteria decision data; Step S43: Optimizing the fire protection implementation rule multi-criteria decision data based on a preset reinforcement learning algorithm to generate optimized fire protection implementation rule multi-criteria decision data; Step S44: Establishing a fire protection implementation rule adaptive control engine based on the optimized fire protection implementation rule multi-criteria decision data.

7. The fire data processing method according to claim 1, characterized in that: Step S5 includes the following steps: Step S51: transmitting the fire risk spatiotemporal assessment index to the fire implementation rule adaptive control engine for multi-agent fire intelligent control parameter analysis to generate multi-agent fire intelligent control parameters; Step S52: Perform collaborative decision optimization analysis on the multi-agent fire protection intelligent control parameters based on collaborative learning technology to generate multi-agent collaborative fire protection intelligent control parameters; Step S53: Execute the multi-agent collaborative fire intelligent control operation according to the collaborative multi-agent fire intelligent control parameters.

8. A smart fire protection platform, characterized in that: Used to execute the fire data processing method according to claim 1, the fire data processing method comprising: The intelligent fire supervision module is used to obtain fire supervision area data; collect fire supervision data from the fire supervision area data to generate fire supervision data; analyze fire supervision anomaly characteristics based on the fire supervision data to generate fire supervision anomaly characteristic data; The fire anomaly analysis module is used to perform data decoupling processing on the fire supervision anomaly feature data to generate decoupled fire supervision anomaly feature data; extract fire anomaly features based on the decoupled fire supervision anomaly feature data to generate fire anomaly feature data; The fire risk spatiotemporal assessment module is used to perform digital spatial modeling of the fire supervision area based on the fire supervision area data to generate a fire supervision area model; transmit the fire anomaly characteristic data to the fire supervision area model to perform simulated fire anomaly spatiotemporal evolution analysis in the fire supervision area to generate simulated fire anomaly spatiotemporal evolution data; and perform spatiotemporal node evaluation of the fire risk index on the simulated fire anomaly spatiotemporal evolution data to generate a fire risk spatiotemporal assessment index; The fire protection implementation rule decision analysis module is used to obtain fire protection implementation rule decisions; perform fire protection implementation rule multi-criteria decision optimization analysis based on the fire protection implementation rule decisions to generate optimized fire protection implementation rule multi-criteria decision data; and establish a fire protection implementation rule adaptive control engine based on the optimized fire protection implementation rule multi-criteria decision data; The multi-agent collaborative firefighting module is used to transmit the fire risk spatiotemporal assessment index to the fire implementation rule adaptive control engine to perform multi-agent collaborative firefighting intelligent control parameter analysis, generate multi-agent collaborative firefighting intelligent control parameters; and execute multi-agent collaborative firefighting intelligent control operations based on the collaborative multi-agent firefighting intelligent control parameters.

Images