Method and system for evaluating fire safety situation based on AI multi-mode perception

Abstract

The invention discloses a fire safety situation assessment method and system based on AI multi-mode perception, and relates to the related field of data processing, and the method comprises the steps: carrying out the regional structure data collection of a target region, and building a calibration regional structure data set; carrying out regional twin modeling, and constructing a basic model; activating a multi-mode sensing layer, performing sensor group communication, and establishing a sensor communication data set; after data preprocessing is carried out on the sensor communication data set, the sensor communication data set is updated to the basic model, multi-modal data authentication is carried out, fire development fitting is carried out according to a multi-modal data authentication result, and a fitting result is established; and performing channel security situation evaluation according to the fitting result, and reporting an evaluation early warning based on the channel security situation evaluation result and the fitting result. The technical problem of emergency response lag caused by insufficient early warning accuracy in existing fire safety situation assessment is solved, and the technical effects of improving the early warning accuracy and improving the emergency response speed are achieved.

Description

Technical Field

[0001] The present application relates to data processing related fields, and in particular, to a fire safety situation assessment method and system based on AI multimodal perception. Background Art

[0002] With the acceleration of urbanization and the increase in high-rise buildings, fire safety issues have become increasingly prominent. Traditional fire safety assessment methods mainly rely on manual inspections and single sensor monitoring. However, manual inspection methods are highly subjective and inefficient. Monitoring fire risks by installing single sensors such as smoke detectors and temperature sensors can only reflect certain aspects of fire characteristics and cannot fully and accurately assess the fire situation, resulting in delayed emergency response and failure to provide early warnings and proactively prevent fires.

[0003] Among the current relevant technologies, fire safety situation assessment has the technical problem of insufficient early warning accuracy, which leads to delayed emergency response. Summary of the invention

[0004] The present application provides a fire safety situation assessment method and system based on AI multimodal perception, collects structural data of the target area, establishes a calibration data set, and performs regional twin modeling based on this data, builds a basic model, activates the multimodal perception layer, collects data from built-in and additional sensors through sensor group communication, forms a sensor communication data set, pre-processes the sensor data set, integrates the processing results into the basic model, uses the updated model to perform multimodal data authentication, fits the fire development based on the authentication results, establishes a fire situation fitting result, evaluates the channel safety situation based on the fitting results, and comprehensively combines the fitting results with the situation evaluation results to issue fire safety assessment warnings and other technical means, thereby achieving the technical effect of improving the accuracy of warnings and thus improving the speed of emergency response.

[0005] The present application provides a fire safety situation assessment method based on AI multimodal perception, including: collecting regional structural data of a target area to establish a calibrated regional structural data set; using the calibrated regional structural data set to perform regional twin modeling and construct a basic model; activating a multimodal perception layer, using the multimodal perception layer to perform sensor group communication, and establishing a sensor communication data set, wherein the sensor communication data set includes built-in sensor communication data and additional sensor communication data of the target area; after data preprocessing the sensor communication data set, updating the data preprocessing result to the basic model, using the updated basic model to perform multimodal data authentication based on an intelligent perception network, performing fire development fitting based on the multimodal data authentication result, and establishing a fitting result; evaluating the channel safety situation based on the fitting result, and reporting an assessment warning based on the channel safety situation evaluation result and the fitting result.

[0006] In a possible implementation, the updated basic model is used to perform multimodal data authentication based on an intelligent perception network, and the following processing is performed: activating a joint association subnetwork within the intelligent perception network, using the joint association subnetwork to perform association authentication based on the calibration area structure dataset and the position coordinates of the sensor group, and establishing a joint association; synchronizing the joint association to a data authentication subnetwork, using the data authentication subnetwork to perform sensor communication dataset authentication analysis based on the joint association, and establishing a multimodal data authentication result.

[0007] In a possible implementation, the data authentication subnetwork is used to perform sensor communication data set authentication analysis based on joint association, and the following processing is performed: determining whether the sensor communication data set contains signal abnormality data of the built-in sensor in the target area; if there is signal abnormality data, tracing back to the abnormal sensor corresponding to the signal abnormality data, and establishing an abnormal communication path; performing abnormal joint authentication of the abnormal sensor and the abnormal communication path based on the sensor communication data set, and establishing an abnormal joint authentication result; constructing location hazard data based on the abnormal joint authentication result, and adding the location hazard data to the multimodal data authentication result.

[0008] In a possible implementation, the fire development fitting is performed based on the multimodal data authentication results, and the following processing is performed: a combustible material distribution is established based on the calibration area structure data set; basic information of the combustible materials is obtained, and a combustible risk with a coordinate identifier is established based on the basic information and the combustible material distribution; the multimodal data authentication results are aligned using the coordinate identifiers, and a joint hazard trigger prediction is performed on the combustible risk using the aligned multimodal data authentication results to establish a joint hazard trigger prediction result; and the fire development fitting is completed using the joint hazard trigger prediction result.

[0009] In a possible implementation, the following processing is also performed: continuous perception monitoring is performed based on the multimodal perception layer to establish a time series perception monitoring result; the channel safety situation evaluation result is verified for consistency using the time series perception monitoring result to establish prediction feedback; self-optimization constraints are established through the prediction feedback, and the fitting optimization of the fire development fitting is performed through the self-optimization constraints.

[0010] In a possible implementation, the following processing is also performed: based on the fitting results, a global analysis of the target area is performed to establish regional focus; based on the regional focus, an additional acquisition distribution is reconstructed, and the real-time position of the additional sensor is obtained; based on the real-time position and the additional acquisition distribution, mobile optimization of the additional sensor is performed to establish a mobile optimization result; and based on the mobile optimization result, mobile management of the additional sensor is performed.

[0011] In a possible implementation, the mobile optimization of the additional sensor is performed according to the real-time position and the additional acquisition distribution, and the following processing is performed: a taboo interval is established based on the fitting result, and the taboo interval is a prohibited interval for movement; a taboo constraint is constructed using the taboo interval, and the mobile optimization is completed based on the taboo constraint.

[0012] The present application also provides an assessment system for fire safety situation based on AI multimodal perception, including: a regional structure data acquisition module, the regional structure data acquisition module is used to collect regional structure data of the target area and establish a calibrated regional structure data set; a regional twin modeling module, the regional twin modeling module is used to use the calibrated regional structure data set to perform regional twin modeling and construct a basic model; a sensor group communication module, the sensor group communication module is used to activate the multimodal perception layer, use the multimodal perception layer to perform sensor group communication, and establish a sensor communication data set, the sensor communication data set includes built-in sensor communication data and additional sensor communication data of the target area; a multimodal data authentication module, the multimodal data authentication module is used to perform data preprocessing on the sensor communication data set, update the data preprocessing result to the basic model, use the updated basic model to perform multimodal data authentication based on the intelligent perception network, perform fire development fitting according to the multimodal data authentication result, and establish a fitting result; an assessment and early warning output module, the assessment and early warning output module is used to evaluate the channel safety situation according to the fitting result, and issue an assessment warning based on the channel safety situation evaluation result and the fitting result.

[0013] The fire safety situation assessment method and system based on AI multimodal perception proposed in this application first collects regional structural data of the target area to establish a calibrated regional structural data set, then uses the calibrated regional structural data set to perform regional twin modeling, constructs a basic model, and then activates the multimodal perception layer, uses the multimodal perception layer to perform sensor group communication, and establishes a sensor communication data set, wherein the sensor communication data set includes built-in sensor communication data and additional sensor communication data of the target area. After data preprocessing of the sensor communication data set, the data preprocessing results are updated to the basic model, and the updated basic model is used to perform multimodal data authentication based on the intelligent perception network, and fire development fitting is performed according to the multimodal data authentication results to establish a fitting result. Finally, the channel safety situation is evaluated according to the fitting result, and an assessment warning is issued based on the channel safety situation evaluation result and the fitting result, thereby achieving the technical effect of improving the accuracy of the warning and thus improving the speed of emergency response. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solution of the embodiment of the present invention, the accompanying drawings of the embodiment of the present invention will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the system according to the embodiment of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.

[0015] Figure 1 A flowchart of a method for evaluating fire safety situation based on AI multimodal perception provided in an embodiment of the present application.

[0016] Figure 2 A schematic diagram of the structure of a fire safety situation assessment system based on AI multimodal perception provided in an embodiment of the present application.

[0017] Explanation of the accompanying drawings: regional structure data acquisition module 10, regional twin modeling module 20, sensor group communication module 30, multimodal data authentication module 40, evaluation and early warning module 50. DETAILED DESCRIPTION

[0018] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0019] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0020] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict. The terms "including" and "having" and any variations are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field to which this application belongs. The terms used herein are for the purpose of describing the embodiments of the present application only.

[0021] The present application embodiment provides a fire safety situation assessment method based on AI multimodal perception, such as Figure 1 As shown, the method includes: Step S100, collect regional structural data of the target area and establish a calibration regional structural data set. Specifically, the target area is an area that needs to be evaluated for fire safety situation, such as buildings, factories, warehouses, etc., and professional data acquisition equipment is used to collect comprehensive structural data of the target area, including the spatial layout of the area, building materials, fire passages, fire equipment location and other information, and integrate this information into a complete calibration regional structural data set, that is, the calibration regional structural data set is a collection of all structural data of the target area, which is used for subsequent modeling and analysis.

[0022] Step S200, using the calibrated regional structure data set to perform regional twin modeling and construct a basic model. Specifically, based on the calibrated regional structure data set, a virtual model of the target area is constructed using digital twin technology. This model maps the physical space of the real world into the digital world for simulation and analysis.

[0023] Step S300, activate the multimodal perception layer, use the multimodal perception layer to perform sensor group communication, and establish a sensor communication data set, which includes the communication data of the built-in sensors in the target area and the communication data of the additional sensors. Specifically, multiple types of sensors (such as temperature sensors, smoke sensors, infrared sensors, etc.) are arranged in the target area to form a multimodal perception layer. These sensors are interconnected through a communication protocol to collect various data in the target area in real time, and these data are integrated into a sensor communication data set. Among them, the built-in sensors in the target area are pre-installed in the target area, and their positions and types are determined according to the characteristics of the target area and the fire safety requirements. The built-in sensors can continuously monitor the status of the target area and send the collected data to the multimodal perception layer for processing through the communication protocol. These data include key parameters such as temperature, humidity, and smoke concentration, which play an important role in timely discovering fire hazards. Additional sensors are sensors added in addition to built-in sensors to obtain more comprehensive information. These additional sensors can be deployed according to specific circumstances, for example, adding temperature sensors in areas with fire hazards, or adding smoke sensors in areas with dense personnel. The data from the additional sensors are also sent to the multimodal perception layer through the communication protocol for processing and integrated with the data from the built-in sensors to form a complete sensor communication data set.

[0024] Step S400, after data preprocessing of the sensor communication data set, the data preprocessing result is updated to the basic model, and the updated basic model is used to perform multimodal data authentication based on the intelligent perception network, and the fire development is fitted according to the multimodal data authentication result to establish the fitting result. Specifically, the sensor communication data set is preprocessed by cleaning, denoising and other preprocessing operations to improve the accuracy and reliability of the data. The preprocessed data is updated to the basic model so that the model can reflect the real-time status of the target area. The multimodal data is authenticated by using an intelligent perception network (a network built using technologies such as deep learning that can process and analyze multimodal data) to ensure the authenticity and validity of the data. Based on the authenticated data, the model is used to simulate and predict the development of the fire, and a fitting result of the fire development is established.

[0025] In a possible implementation, the updated basic model is used to perform multimodal data authentication based on the intelligent perception network, and step S400 further includes step S410, activating the joint association sub-network in the intelligent perception network, using the joint association sub-network to perform association authentication based on the calibration area structure data set and the position coordinates of the sensor group, and establishing a joint association. Specifically, in the intelligent perception network, there is a sub-network that is specifically responsible for processing the association relationship between different data sources, namely the joint association sub-network. Start this sub-network and put it in a working state. The joint association sub-network obtains the calibration area structure data set (including the physical structure information of the target area) and the position coordinates of the sensor group (indicating the specific position of each sensor in the target area). Based on the calibration area structure data set and the position coordinates of the sensor group, the joint association sub-network performs a series of algorithm operations to establish the association between these sensors and the physical structure of the target area. This association is the basis for subsequent data authentication, ensuring that the accuracy and reliability of the sensor data can match its position in the physical space.

[0026] Step S420, synchronize the joint association to the data authentication subnetwork, use the data authentication subnetwork to perform authentication analysis of the sensor communication data set based on the joint association, and establish a multimodal data authentication result. Specifically, the joint association information established in step S410 is synchronized to the data authentication subnetwork, which is another subnetwork in the intelligent perception network and is specifically responsible for processing the authentication analysis of sensor data. The data authentication subnetwork uses the synchronized joint association information to perform authentication analysis on the sensor communication data set, including checks on the integrity, accuracy, consistency, etc. of the data to ensure that the sensor data is reliable and valid. Based on the results of the authentication analysis, the data authentication subnetwork generates a multimodal data authentication result, which is an authentication report that integrates information from multiple sensor data sources and reflects the perception accuracy of the fire situation in the target area. This implementation method ensures the accuracy and reliability of sensor data through the processing of the joint association subnetwork and the data authentication subnetwork, reduces false alarms and missed alarms, and improves the accuracy of fire situation assessment.

[0027] In a possible implementation, the sensor communication data set authentication analysis based on joint association is performed using the data authentication subnetwork, and step S420 further includes step S421, determining whether the sensor communication data set contains signal abnormality data of the built-in sensor in the target area. Specifically, the data authentication subnetwork first performs a comprehensive scan and detection on the sensor communication data set. During the scanning process, any data that does not conform to the normal signal pattern of the built-in sensor in the target area is identified, and these data are regarded as signal abnormality data. For example, if the signal of a sensor is suddenly interrupted or becomes extremely unstable, this indicates that there may be some kind of problem. Once the abnormal data is identified, it is immediately marked for subsequent analysis and processing. Step S422, if there is signal abnormality data, the abnormal sensor corresponding to the signal abnormality data is traced back and an abnormal communication path is established. Specifically, based on the abnormal data marked in step S421, the specific abnormal sensor is traced back and located. The communication path between the abnormal sensor and the system component is analyzed to determine whether there are any potential faults or interference. Through comprehensive analysis, an abnormal communication path is established, including abnormal sensors, transmission lines, relay equipment, etc. Step S423, based on the sensor communication data set, perform abnormal joint authentication of the abnormal sensor and the abnormal communication path, and establish an abnormal joint authentication result. Specifically, integrate all data related to the abnormal sensor and the abnormal communication path in the sensor communication data set. Use an algorithm or model to perform joint authentication analysis on these data to further confirm the authenticity of the abnormal sensor and the abnormal communication path. Based on the analysis results, establish an abnormal joint authentication result, which describes the abnormal sensor, the abnormal communication path and the association between them. Step S424, construct position hazard data according to the abnormal joint authentication result, and add the position hazard data to the multimodal data authentication result. Specifically, based on the abnormal joint authentication result, construct a position hazard data set, which describes the location of the abnormal sensor and the abnormal communication path and the dangers they may bring. Integrate and update this position hazard data set into the multimodal data authentication result to provide comprehensive fire situation assessment information. This implementation method can more accurately identify potential fire risk points by analyzing abnormal sensors and abnormal communication paths, thereby improving the accuracy of fire situation assessment.

[0028] In a possible implementation, the fire development fitting is performed according to the multimodal data authentication result, and step S400 further includes step S430, establishing the distribution of combustibles based on the calibration area structure data set. Specifically, information about the spatial structure of buildings, rooms, corridors, etc. is extracted from the calibration area structure data set, and the combustibles in the calibration area are identified and marked according to the spatial structure information, combined with possible combustible types (such as wood, paper, plastic, etc.) and distribution patterns. These combustibles are presented on the distribution map in a graphical manner to form a combustible distribution map. This distribution map shows the location and quantity of combustibles in the target area. Step S440, obtain basic information of combustibles, and establish a combustible risk with coordinate identification according to the basic information and the combustible distribution. Specifically, basic information of combustibles, such as type, burning speed, calorific value, etc., is collected through a database. Based on the basic information of combustibles and the combustible distribution map, the risk value of each combustible point or area is calculated, and this risk value integrates factors such as the type, quantity, location and possible fire source of the combustibles. In order to correspond the flammable risk to the actual location of the target area, a coordinate identifier is assigned to each flammable risk point or area.

[0029] Step S450, aligning the multimodal data authentication results using the coordinate identifier, and using the aligned multimodal data authentication results to predict the combustible risk for joint hazard triggering, and establishing a joint hazard triggering prediction result. Specifically, according to the coordinate identifier, each data point in the multimodal data authentication result is matched with the combustible risk to ensure the consistency of the data in space. By comparing and adjusting the timestamp, the temporal consistency of the multimodal data authentication result and the combustible risk is ensured. The aligned multimodal data authentication results are fused with the combustible risk, that is, data of different sources and different types (such as temperature, smoke concentration, flame detection, etc.) are integrated together to form a more comprehensive fire situation perception, and based on the fused data, a deep learning model is used to predict hazard triggering. This model comprehensively considers the risk value of combustibles, environmental conditions (such as temperature, humidity, etc.) and abnormal information in the multimodal data authentication results (such as signal loss, abnormal sensors, etc.) to predict the possibility and location of fire. Step S460, completing the fire development fitting with the joint hazard triggering prediction result. Specifically, the joint hazard trigger prediction results are embedded into the dynamic model of fire development to simulate the spread of fire. Based on the integrated data, the fire development is fitted, including the simulation of the fire spread speed, direction, and possible impact range. This implementation method conducts joint hazard trigger prediction by integrating the distribution of combustibles, basic information, and multimodal data authentication results, more accurately assessing the fire situation and improving the accuracy of fire situation assessment.

[0030] Step S500, the safety situation of the channel is evaluated according to the fitting result, and an evaluation warning is issued based on the evaluation result of the safety situation of the channel and the fitting result. Specifically, based on the fitting result of the fire development, the safety situation of the fire channel, evacuation path, etc. in the target area is evaluated. According to the evaluation result and the fitting result, the evaluation warning information is generated, such as the fire risk level, possible evacuation path, etc. The evaluation warning information is promptly conveyed to the relevant personnel or system so that necessary countermeasures can be taken. The embodiment of the present application adopts the acquisition of structural data of the target area, establishes a calibration data set, and performs regional twin modeling based on it, constructs a basic model, activates the multimodal perception layer, collects data from built-in and additional sensors through sensor group communication, forms a sensor communication data set, pre-processes the sensor data set, integrates the processing results into the basic model, uses the updated model to perform multimodal data authentication, performs fire development fitting based on the authentication result, establishes the fire situation fitting result, evaluates the channel safety situation according to the fitting result, and comprehensively combines the fitting result with the situation evaluation result, and issues the fire safety evaluation warning and other technical means, so as to achieve the technical effect of improving the accuracy of the warning and thus improving the emergency response speed.

[0031] In a possible implementation, the method further includes step S600, performing continuous perception monitoring based on the multimodal perception layer to establish a time series perception monitoring result. Specifically, based on step S300, the multimodal perception layer continuously collects data from built-in sensors and additional sensors in the target area, including but not limited to temperature, smoke concentration, flame detection, sound (such as explosion or cry for help) and other multi-modal information. The collected multimodal data will be integrated and preprocessed to eliminate noise, fill missing values, and perform necessary format conversion to ensure data consistency and accuracy. The processed data is arranged in chronological order to form a time series data set. Using time series analysis methods, such as time series analysis, sliding window technology, etc., the time series data set is further analyzed and mined to extract key features and trends related to fire development. Based on the results of the time series data analysis, time series perception monitoring results are generated, which reflect the changes in the fire safety situation in the target area over a continuous time.

[0032] Step S700, using the time series perception monitoring results to verify the consistency of the channel safety situation evaluation results, and establish prediction feedback. Specifically, the time series perception monitoring results are compared and analyzed with the channel safety situation evaluation results generated in step S500 to verify the consistency and difference between the two, including comparing whether the predicted trend of fire development is consistent with the actual monitored trend, and whether there are significant deviations or anomalies. Based on the results of consistency verification, generate prediction feedback. If the prediction results are highly consistent with the actual monitoring results, it indicates that the prediction model has a high accuracy; if there is a significant deviation, it is necessary to adjust the prediction model or re-evaluate the quality of the data. Step S800, establish a self-optimization constraint through the prediction feedback, and optimize the fitting of the fire development fitting through the self-optimization constraint. Specifically, according to the prediction feedback, formulate self-optimization constraints, including threshold requirements for data quality, adjustment rules for the prediction model or optimization range of algorithm parameters, etc. Under the guidance of the self-optimization constraints, the process of fire development fitting is optimized, including retraining the prediction model, adjusting the algorithm parameters, improving the data preprocessing steps, etc. The optimized fire development fitting results will be compared and analyzed again with the actual monitoring results to evaluate the effect of the optimization. If the optimized results are more consistent with the actual monitoring results, it means that the optimization measures are effective; otherwise, the self-optimization constraints need to be further adjusted. This implementation method can timely detect the differences between the prediction model and the actual monitoring results through continuous perception monitoring and consistency verification, and perform fitting optimization through self-optimization constraints, thereby improving the prediction accuracy of fire development, enhancing the robustness and stability of the system, and providing more reliable technical support for fire safety.

[0033] In a possible implementation, the method further includes step S900, performing a global analysis of the target area based on the fitting results and establishing a regional focus. Specifically, the fire development fitting result obtained in step S400 is integrated with the channel safety situation evaluation result in step S500 to form a comprehensive fire safety situation data set. The integrated data set is globally analyzed using an AI algorithm to identify areas with higher fire risks in the target area, namely "regional concerns". These areas can be critical paths for the spread of fire, densely populated areas, or storage areas for flammable items. Based on the results of the global analysis, the specific regional focus range is determined and marked on the map of the target area for subsequent targeted monitoring and management.

[0034] Step S1000, reconstructing the additional acquisition distribution according to the regional focus, and obtaining the real-time position of the additional sensor. Specifically, based on the regional focus determined in step S900, re-planning the acquisition distribution of the additional sensor, including determining the number, type, location and monitoring frequency of the additional sensor. Using positioning technology (such as GPS, Wi-Fi positioning, etc.) to obtain the real-time location information of the current additional sensor, this information is used for the subsequent mobile optimization of the additional sensor. Step S1100, performing mobile optimization of the additional sensor according to the real-time position and the additional acquisition distribution, and establishing a mobile optimization result. Specifically, using a mobile optimization algorithm (such as a path planning algorithm, etc.) to optimize the real-time position and the additional acquisition distribution of the additional sensor, including determining the optimal mobile path, the stop position and other parameters of the additional sensor. Based on the calculation results of the mobile optimization algorithm, the mobile optimization result of the additional sensor is established to guide the subsequent mobile management of the additional sensor. Step S1200, performing mobile management of the additional sensor according to the mobile optimization result. Specifically, the mobile optimization result obtained in step S1100 is converted into specific mobile instructions and sent to the additional sensors. These instructions include specific information such as the moving path and the stop position. The additional sensors move according to the mobile instructions and continuously collect and transmit data during the movement. At the same time, the mobile status and collected data of the additional sensors are monitored in real time to ensure the smooth progress of mobile management. According to the mobile status and collected data of the additional sensors, the effect of mobile management is evaluated and feedback is given. If the mobile management effect is not good or there are abnormal conditions, the mobile instructions and additional collection distribution are adjusted in time to improve the monitoring efficiency. This implementation method accurately determines the high-risk areas in the target area through global analysis and establishment of regional concerns, so as to carry out targeted reconstruction of the collection distribution and mobile optimization of the additional sensors, which not only improves the monitoring efficiency, but also reduces unnecessary waste of resources.

[0035] In a possible implementation, the mobile optimization of the additional sensor is performed according to the real-time position and the additional acquisition distribution, and step S1100 further includes step S1110, establishing a taboo interval based on the fitting result, and the taboo interval is a prohibited interval for movement. Specifically, the fire development fitting result obtained in step S400 is deeply analyzed to identify the key path of fire spread, high-risk areas, and possible fire source locations. Based on the analysis of the fitting results, combined with the structural characteristics of the target area, the distribution of personnel, the storage of flammable items and other factors, the intervals where the additional sensor is prohibited from moving, namely "taboo intervals", are determined. These intervals are areas where the fire is extremely easy to spread, areas where the personnel are dense and difficult to evacuate, or areas where key facilities that are crucial to fire safety are located, which are used to prevent the sensor from interfering with fire control or causing personal injury during movement. Step S1120, using the taboo interval to construct a taboo constraint, and completing the mobile optimization based on the taboo constraint. Specifically, the taboo interval determined in step S1110 is converted into a specific taboo constraint condition, which is used to limit the possible path of the additional sensor in the mobile optimization process. By modifying the logic of the algorithm or adding additional judgment conditions to the algorithm, taboo constraints are introduced into the mobile optimization algorithm to ensure that the additional sensor will not enter the taboo interval during the movement. Under the restriction of the taboo constraint, the mobile optimization algorithm is run to calculate the optimal movement path and position of the additional sensor, which is used to guide the subsequent movement management of the additional sensor. This implementation method ensures that the additional sensor will not enter the area where the fire is easy to spread or the people are crowded during the movement by setting the taboo interval and taboo constraint, so as to avoid the sensor becoming an obstacle to the spread of the fire or causing harm to people.

[0036] In the above, refer to Figure 1 The fire safety situation assessment method based on AI multimodal perception according to an embodiment of the present invention is described in detail. Figure 2 The present invention describes an evaluation system for fire safety situation based on AI multimodal perception according to an embodiment of the present invention.

[0037] The fire safety situation assessment system based on AI multimodal perception according to the embodiment of the present invention is used to solve the technical problem of insufficient warning accuracy in the existing fire safety situation assessment, which leads to delayed emergency response, and achieves the technical effect of improving the accuracy of warning and thus improving the speed of emergency response. The fire safety situation assessment system based on AI multimodal perception includes: a regional structure data acquisition module 10, a regional twin modeling module 20, a sensor group communication module 30, a multimodal data authentication module 40, and an assessment early warning module 50.

[0038] The regional structure data acquisition module 10 is used to collect regional structure data of the target area and establish a calibrated regional structure data set; the regional twin modeling module 20 is used to use the calibrated regional structure data set to perform regional twin modeling and construct a basic model; the sensor group communication module 30 is used to activate the multimodal perception layer, use the multimodal perception layer to perform sensor group communication, and establish a sensor communication data set, wherein the sensor communication data set includes built-in sensor communication data and additional sensor communication data of the target area; the multimodal data authentication module 40 is used to perform data preprocessing on the sensor communication data set, update the data preprocessing result to the basic model, use the updated basic model to perform multimodal data authentication based on the intelligent perception network, perform fire development fitting according to the multimodal data authentication result, and establish a fitting result; the evaluation early warning output module 50 is used to evaluate the channel safety situation according to the fitting result, and issue an evaluation warning based on the channel safety situation evaluation result and the fitting result.

[0039] The specific configuration of the multimodal data authentication module 40 will be described in detail below. As described above, the updated basic model is used to perform multimodal data authentication based on the intelligent sensing network, and the multimodal data authentication module 40 may further include: an association authentication unit for activating a joint association subnetwork within the intelligent sensing network, using the joint association subnetwork to perform association authentication based on the calibration area structure data set and the position coordinates of the sensor group, and establish a joint association; an authentication analysis unit for synchronizing the joint association to the data authentication subnetwork, using the data authentication subnetwork to perform authentication analysis of the sensor communication data set based on the joint association, and establish a multimodal data authentication result.

[0040] Among them, the sensor communication data set authentication analysis based on joint association is performed using the data authentication subnetwork, and the authentication analysis unit may further include: a judgment subunit is used to judge whether the sensor communication data set contains signal abnormality data of the built-in sensor in the target area; an abnormal communication path establishment subunit is used to trace back the abnormal sensor corresponding to the signal abnormality data if there is signal abnormality data, and establish an abnormal communication path; an abnormal joint authentication subunit is used to perform abnormal joint authentication of the abnormal sensor and the abnormal communication path based on the sensor communication data set, and establish an abnormal joint authentication result; a location hazard data construction subunit is used to construct location hazard data according to the abnormal joint authentication result, and add the location hazard data to the multimodal data authentication result.

[0041] Among them, the fire development fitting is performed according to the multimodal data authentication result, and the multimodal data authentication module 40 may further include: a combustible distribution establishment unit is used to establish the combustible distribution based on the calibration area structure data set; a combustible risk establishment unit is used to obtain basic information of combustibles, and establish a combustible risk with coordinate identification according to the basic information and the combustible distribution; a joint hazard trigger prediction unit is used to align the multimodal data authentication result with the coordinate identification, and use the aligned multimodal data authentication result to perform a joint hazard trigger prediction on the combustible risk, and establish a joint hazard trigger prediction result; a fire development fitting unit is used to complete the fire development fitting with the joint hazard trigger prediction result.

[0042] Among them, the fire safety situation assessment system based on AI multimodal perception can further include: a continuous perception monitoring module is used to perform continuous perception monitoring based on the multimodal perception layer, and establish a time series perception monitoring result; a consistency verification module is used to use the time series perception monitoring result to perform consistency verification on the channel safety situation evaluation result, and establish prediction feedback; a fitting optimization module is used to establish a self-optimization constraint through the prediction feedback, and perform fitting optimization of the fire development fitting through the self-optimization constraint.

[0043] Among them, the fire safety situation assessment system based on AI multimodal perception can further include: a global analysis module for performing a global analysis of the target area based on the fitting results and establishing regional focus; an additional acquisition distribution reconstruction module for reconstructing the additional acquisition distribution according to the regional focus and obtaining the real-time position of the additional sensor; an additional sensor mobile optimization module for performing mobile optimization of the additional sensor according to the real-time position and the additional acquisition distribution, and establishing a mobile optimization result; an additional sensor mobile management module for performing additional sensor mobile management according to the mobile optimization result.

[0044] Among them, the mobile optimization of the additional sensor is performed according to the real-time position and the additional collection distribution, and the mobile optimization module of the additional sensor can further include: a taboo interval establishment unit is used to establish a taboo interval based on the fitting result, and the taboo interval is a prohibited interval for movement; a taboo constraint construction unit is used to construct a taboo constraint using the taboo interval, and complete the mobile optimization based on the taboo constraint.

[0045] The fire safety situation assessment system based on AI multimodal perception provided by the embodiment of the present invention can execute the fire safety situation assessment method based on AI multimodal perception provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0046] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules may be used and run on the user terminal and / or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0047] The above specific implementation manner does not constitute a limitation to the protection scope of the present application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application. In some cases, the actions or steps recorded in the present application can be performed in an order different from that in the embodiment and can still achieve the desired results. In addition, the process depicted in the accompanying drawings does not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. The fire safety situation assessment method based on AI multimodal perception is characterized by: The method comprises: Collect regional structural data of the target area and establish a calibration regional structural data set; Using the calibrated regional structure data set to perform regional twin modeling and construct a basic model; activating a multimodal perception layer, utilizing the multimodal perception layer for sensor group communication, and establishing a sensor communication data set, wherein the sensor communication data set includes communication data of built-in sensors in a target area and communication data of additional sensors; After data preprocessing is performed on the sensor communication data set, the data preprocessing result is updated to the basic model, the updated basic model is used to perform multimodal data authentication based on the intelligent perception network, and the fire development fitting is performed according to the multimodal data authentication result to establish the fitting result; Performing a channel security situation evaluation according to the fitting result, and reporting an evaluation warning based on the channel security situation evaluation result and the fitting result; The method of using the updated basic model to perform multimodal data authentication based on an intelligent perception network includes: activating a joint association subnetwork within the intelligent sensing network, using the joint association subnetwork to perform association authentication based on the calibration area structure data set and the position coordinates of the sensor group, and establishing a joint association; Synchronizing the joint association to a data authentication subnetwork, using the data authentication subnetwork to perform authentication analysis of sensor communication data sets based on the joint association, and establishing a multimodal data authentication result; The fire development fitting according to the multimodal data authentication result includes: Establishing a combustible distribution based on the calibration area structure data set; Obtaining basic information of combustibles, and establishing combustible risks with coordinate identifiers based on the basic information and the distribution of the combustibles; Aligning the multimodal data authentication results by using the coordinate identifiers, performing a joint hazard trigger prediction on the flammable risk by using the aligned multimodal data authentication results, and establishing a joint hazard trigger prediction result; The fire development fitting is completed with the combined hazard trigger prediction result.

2. The fire safety situation assessment method based on AI multimodal perception as claimed in claim 1, characterized in that: The using the data authentication subnetwork to perform sensor communication data set authentication analysis based on joint association includes: Determine whether the sensor communication data set contains abnormal signal data of a built-in sensor in the target area; If there is abnormal signal data, trace back to the abnormal sensor corresponding to the abnormal signal data and establish an abnormal communication path; Performing abnormal joint authentication of the abnormal sensor and the abnormal communication path based on the sensor communication data set, and establishing an abnormal joint authentication result; Construct location danger data according to the abnormal joint authentication result, and add the location danger data to the multimodal data authentication result.

3. The fire safety situation assessment method based on AI multimodal perception as claimed in claim 1, characterized in that: The method further comprises: Perform continuous perception monitoring based on the multimodal perception layer and establish time-series perception monitoring results; Using the time series perception monitoring results to verify the consistency of the channel security situation evaluation results, and establish prediction feedback; A self-optimization constraint is established through the prediction feedback, and the fitting optimization of the fire development fitting is performed through the self-optimization constraint.

4. The fire safety situation assessment method based on AI multimodal perception as claimed in claim 1, characterized in that: The method further comprises: Performing a global analysis of the target area based on the fitting results and establishing regional attention; Reconstructing the additional acquisition distribution according to the regional focus and obtaining the real-time position of the additional sensor; Perform mobile optimization of the additional sensor according to the real-time position and the additional acquisition distribution, and establish a mobile optimization result; Additional sensor movement management is performed according to the movement optimization result.

5. The fire safety situation assessment method based on AI multimodal perception as claimed in claim 4 is characterized in that: The mobile optimization of the additional sensor according to the real-time position and the additional acquisition distribution includes: Establishing a taboo interval based on the fitting result, wherein the taboo interval is a prohibited interval of movement; The taboo interval is used to construct a taboo constraint, and mobile optimization is completed based on the taboo constraint.

6. Fire safety situation assessment system based on AI multimodal perception, characterized by: The system is used to implement the fire safety situation assessment method based on AI multimodal perception according to any one of claims 1 to 5, and the system includes: A regional structure data acquisition module, wherein the regional structure data acquisition module is used to collect regional structure data of the target area and establish a calibration regional structure data set; A regional twin modeling module, wherein the regional twin modeling module is used to perform regional twin modeling using the calibrated regional structure data set to construct a basic model; A sensor group communication module, the sensor group communication module is used to activate the multimodal perception layer, use the multimodal perception layer to perform sensor group communication, and establish a sensor communication data set, the sensor communication data set includes target area built-in sensor communication data and additional sensor communication data; A multimodal data authentication module, wherein the multimodal data authentication module is used to perform data preprocessing on the sensor communication data set, update the data preprocessing result to the basic model, perform multimodal data authentication based on the intelligent perception network using the updated basic model, perform fire development fitting according to the multimodal data authentication result, and establish a fitting result; An evaluation and early warning module is used to evaluate the channel security situation according to the fitting result, and to issue an evaluation warning based on the channel security situation evaluation result and the fitting result.

Images