Building fire monitoring system based on Internet of Things

Abstract

The present invention discloses a building fire monitoring system based on the Internet of Things, including a data acquisition module, a building fire assessment model building module and a building fire monitoring module. The present invention belongs to the field of data processing, and specifically refers to a building fire monitoring system based on the Internet of Things. The scheme is based on multi-scale time series feature extraction and adaptive time step scaling mechanism to achieve synchronous detection of instantaneous fire signals and periodic equipment anomalies; based on time domain feature focusing and nonlinear sensor intensity change and dynamic trend adjustment mechanism, high-precision positioning of fire precursors and equipment failures is achieved; the introduction of secondary screening mechanism and dynamic threshold adjustment strategy significantly improves the accuracy of fire signal and equipment anomaly detection results; by constructing early detection reward loss, the penalty for early exposure of fire precursor signals or equipment failures is reduced, which meets the early warning needs of fire and abnormal equipment failures, and thus comprehensively improves the early warning accuracy and response efficiency of the building fire monitoring system.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular to a building fire monitoring system based on the Internet of Things. Background Art

[0002] The building fire monitoring system is an intelligent system that uses sensors, the Internet of Things, data analysis and other technologies to monitor and manage fire risks, fire equipment status and spatial environment inside buildings in real time. Its core goal is to ensure the fire safety of buildings, improve the efficiency of fire prevention and control, and reduce casualties and property losses through precise monitoring, timely warning and intelligent response. However, traditional fire monitoring and early warning systems are difficult to adapt to the dynamic fluctuations of complex building environments, resulting in low accuracy in fire signal detection, insufficient ability to identify hidden fire precursors and equipment abnormal patterns, and thus difficulty in accurately warning of potential fire risks; traditional fire monitoring and early warning systems are difficult to adapt to the dynamic changes of environmental and equipment data in different time periods, have a high false alarm rate and are severely affected by noise, and the timeliness of fire and equipment abnormality predictions is insufficient, which leads to the problem of limited reliability and stability of the overall monitoring system. Summary of the invention

[0003] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides a building fire monitoring system based on the Internet of Things. In view of the problem that traditional fire monitoring and early warning systems are difficult to adapt to the dynamic fluctuations of complex building environments, resulting in low accuracy in fire signal detection, insufficient ability to identify hidden fire precursors and equipment abnormal patterns, and thus difficult to accurately warn of potential fire risks, this solution is based on multi-scale time series feature extraction and adaptive time step scaling mechanism to achieve synchronous detection of instantaneous fire signals and periodic equipment anomalies, and accurately locate the time point when a fire or abnormal event occurs; based on time domain feature focusing and nonlinear sensor intensity changes and dynamic trend adjustment mechanism, high-precision positioning of fire precursors and equipment failures is achieved, thereby identifying tiny but potentially risky fire signals. The solution introduces a secondary screening mechanism and a dynamic threshold adjustment strategy to significantly improve the accuracy and reliability of fire signal and equipment anomaly detection results. By constructing an early detection reward loss, it reduces the penalty for early exposure of fire precursor signals or equipment failures, meets the early warning needs of fire and abnormal equipment failures, and thus comprehensively improves the early warning accuracy and response efficiency of building fire monitoring systems.

[0004] The technical solution adopted by the present invention is as follows: the building fire monitoring system based on the Internet of Things provided by the present invention includes a data acquisition module, a building fire assessment model establishment module and a building fire monitoring module;

[0005] The data acquisition module collects fire monitoring data of historical buildings;

[0006] The building fire assessment model establishment module realizes the synchronous detection of instantaneous fire signals and periodic equipment anomalies through multi-scale time series feature extraction, adaptive time step scaling mechanism, dual-channel joint feature extraction network and loss function optimization, thereby completing the establishment of the building fire assessment model;

[0007] The building fire monitoring module detects and warns users of real-time data based on the established building fire assessment model.

[0008] Furthermore, in the data acquisition module, the fire monitoring data of the historical building includes environmental monitoring data, equipment operation data, space status data, time and fire assessment results; the fire assessment results are used as data labels; and the collected data are converted and standardized.

[0009] Furthermore, the building fire assessment model establishment module specifically includes the following contents:

[0010] Multi-scale time series feature extraction unit; the operations performed are: sensor data channel downsampling: extract the time step features of sensor data through 1D convolution kernel and reduce the number of channels, and introduce an adaptive time step scaling mechanism to dynamically adjust the time step to capture the instantaneous high-frequency signals of fire events and periodic abnormal signals of equipment failures; time series pooling downsampling: use time pooling to reduce the number of time steps; time series upsampling: restore the time step through inverse convolution to retain the global time series features; multi-scale feature aggregation: directly aggregate the brother feature blocks from different downsampling and upsampling paths; and introduce the time domain feature focusing and dynamic update mechanism to focus on the time point when the abnormal event occurs, highlighting the key time steps of fire and equipment failure; expressed as: ; ; ; ; ; ; ;in, is the original input time series data, t is the time step; is a one-dimensional convolution operation, k is the kernel size; is the sigmoid activation function; It is a one-dimensional pooling operation; and They are channel downsampling and timing upsampling outputs respectively; is the deconvolution operation; k1 is the number of channels; and They are the results of multi-scale feature aggregation and time domain feature focusing; are features from different channels; and They are average pooling and maximum pooling operations respectively; is the final weighted output; is the adaptive time step scaling parameter; is the dynamic weight coefficient; the GRU module is used to capture the dependencies between time steps; and are the upper and lower limits of the time step scaling factor; It is the softplus function; is the rate of change of the current time step; and is the original time series data corresponding to the time step; is a smoothing term;

[0011] Anomaly detection unit; specifically, it includes constructing a dual-channel joint feature extraction network, outputting the sensor intensity change and sensor dynamic trend prediction results respectively through the shared feature extraction layer; reconstructing the time domain signal through ISTFT, calculating the reconstruction error to determine the abnormal fire signal and equipment failure behavior of the corresponding time step; introducing a secondary screening mechanism, for the abnormal points screened out initially, combining the local reconstruction error and historical fire statistical characteristics to calculate the comprehensive anomaly score, and making a final judgment on the abnormal points based on this; expressed as: ; ; ; ; ; ; expressed as: ; ;in, is the predicted intensity of sensor intensity change; is the predicted dynamic trend; SFN(·) is a dual-channel joint feature extraction network; is the reconstructed time domain signal; is the window function; L is the reconstruction window length, and m is the window index; and is the predicted value under the corresponding window; R is the step size of the window movement; Anomaly1 and Anomaly are the initial anomaly judgment and the final anomaly judgment results respectively; 1 is abnormal and 0 is normal; is the input original time domain signal; It is the signal reconstructed by the network of sensor intensity changes and sensor dynamic trends; is the dynamic error threshold; and are the historical mean and standard deviation of the reconstruction error, respectively; is the adjustment factor; and is the weight coefficient; is the anomaly score value; is the local reconstruction error; and They are the number of historical abnormal statistics and the total number of historical statistics; is the abnormal threshold; and is the adjustment parameter; and They are the local changes of sensor intensity change and sensor dynamic trend, respectively;

[0012] Loss function design unit; build early detection reward loss ; expressed as: ; Construct discrete error loss , ensuring that the prediction results are more concentrated on the key time steps; expressed as: ; get the final loss , expressed as: ; ; ; Update the model parameters based on the gradient descent algorithm, and then establish a building fire assessment model; Among them, is the predicted time step of fire and equipment failure; is the actual time step of fire and equipment failure; N is the total number of time steps; , and They are the number of times the predicted fire time step is concentrated after the actual fire time step, the predicted fire time step is concentrated before the actual fire time step, and the predicted fire time step is concentrated in the actual fire time step and completely overlaps; is the root mean square error; and is the loss weight; is the standard deviation of the abnormal time step error; mean(·) is the average value; is the early detection reward loss for the j-th input.

[0013] Furthermore, the building fire monitoring module collects building fire data in real time based on the established building fire assessment model, and monitors the building fire according to the fire assessment results output by the building fire assessment model.

[0014] The beneficial effects achieved by the present invention using the above scheme are as follows:

[0015] (1) In view of the problem that traditional fire monitoring and early warning systems are difficult to adapt to the dynamic fluctuations of complex building environments, resulting in low accuracy in fire signal detection and insufficient ability to identify hidden fire precursors and equipment abnormal patterns, making it difficult to accurately warn of potential fire risks, this solution is based on multi-scale time series feature extraction and adaptive time step scaling mechanism to achieve synchronous detection of instantaneous fire signals and periodic equipment anomalies, and accurately locate the time point when a fire or abnormal event occurs; based on time domain feature focusing and nonlinear sensor intensity changes and dynamic trend adjustment mechanism, high-precision positioning of fire precursors and equipment failures can be achieved, and small but potentially risky fire signals and equipment abnormal patterns can be identified, comprehensively improving the response effect and risk warning capability of building fire monitoring systems.

[0016] (2) In view of the problems that traditional fire monitoring and early warning systems are difficult to adapt to the dynamic changes of environmental and equipment data in different time periods, have a high false alarm rate and are severely affected by noise, and the timeliness of fire and equipment abnormality predictions is insufficient, which in turn limits the reliability and stability of the overall monitoring system, this solution significantly improves the accuracy and reliability of fire signal and equipment abnormality detection results by introducing a secondary screening mechanism and a dynamic threshold adjustment strategy. By constructing an early detection reward loss, it reduces the penalty for early exposure of fire precursor signals or equipment failures, meets the early warning needs of fire and abnormal equipment failures, and thus comprehensively improves the early warning accuracy and response efficiency of the building fire monitoring system. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A schematic diagram of the process of the building fire monitoring system based on the Internet of Things provided by the present invention;

[0018] Figure 2 Schematic diagram of the process for establishing modules for the building fire assessment model.

[0019] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0020] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0021] In the description of the present invention, it should be understood that terms such as “upper”, “lower”, “front”, “back”, “left”, “right”, “top”, “bottom”, “inside” and “outside” indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore should not be understood as limiting the present invention.

[0022] Example 1, see Figure 1 The building fire monitoring system based on the Internet of Things provided by the present invention includes a data acquisition module, a building fire assessment model establishment module and a building fire monitoring module;

[0023] The data acquisition module collects historical building fire monitoring data; and sends the data to the building fire assessment model establishment module;

[0024] The building fire assessment model establishment module realizes the synchronous detection of instantaneous fire signals and periodic equipment anomalies through multi-scale time series feature extraction, adaptive time step scaling mechanism, dual-channel joint feature extraction network and loss function optimization; thereby completing the establishment of the building fire assessment model; and sending the data to the building fire monitoring module;

[0025] The building fire monitoring module detects and warns users of real-time data based on the established building fire assessment model.

[0026] Example 2, see Figure 1 This embodiment is based on the above embodiment. In the data acquisition module, the fire monitoring data of the historical building includes environmental monitoring data, equipment operation data, space status data, time and fire assessment results; the fire assessment results are used as data labels; the fire assessment results include no risk, low risk fire alarm, medium risk fire alarm and high risk fire alarm; and the collected data are converted and standardized; the environmental monitoring data includes temperature, humidity, smoke concentration, carbon monoxide concentration, oxygen concentration and air quality data; the equipment operation data includes fire extinguishing equipment status, electrical equipment status, alarm equipment status, equipment battery power and equipment communication status; the space status data includes door and window status, personnel activity data and monitoring video data.

[0027] Example 3, see Figure 1 and Figure 2 Based on the above embodiment, this embodiment includes the following contents:

[0028] Multi-scale time series feature extraction unit; in order to extract multi-scale time series features for the sensor data of the building fire monitoring system and realize the accurate capture of fire events, feature extraction and aggregation analysis are performed on the input multi-dimensional time series data; the execution operations are: sensor data channel downsampling: extract the time step features of the sensor data through the 1D convolution kernel and reduce the number of channels, and introduce an adaptive time step scaling mechanism to dynamically adjust the time step to capture the instantaneous high-frequency signals of fire events and the periodic abnormal signals of equipment failures; time series pooling downsampling: use time pooling to reduce the number of time steps; time series upsampling: restore the time step through inverse convolution to retain the global time series features; multi-scale feature aggregation: directly aggregate the brother feature blocks from different downsampling and upsampling paths; and introduce the time domain feature focusing and dynamic update mechanism to focus on the time point when the abnormal event occurs, highlighting the key time steps of fire and equipment failure; expressed as: ; ; ; ; ; ; ;in, is the original input time series data, t is the time step; It is a one-dimensional convolution operation, k is the kernel size, and its value range is [1,7]; is the sigmoid activation function; It is a one-dimensional pooling operation; and They are channel downsampling and timing upsampling outputs respectively; is the deconvolution operation; k1 is the number of channels; and They are the results of multi-scale feature aggregation and time domain feature focusing; are features from different channels; and They are average pooling and maximum pooling operations respectively; is the final weighted output; is the adaptive time step scaling parameter; is the dynamic weight coefficient; the GRU module is used to capture the dependencies between time steps; and are the upper and lower limits of the time step scaling factor, with value ranges of [1.5, 3.0] and [0.5, 0.5] respectively; It is the softplus function; is the rate of change of the current time step; and is the original time series data corresponding to the time step; It is a smoothing term; it captures abnormal changes in environmental monitoring data, equipment operation data, and spatial status data at different time scales, realizes synchronous analysis of instantaneous fire signals and periodic equipment failures, and accurately identifies potential threats of low-risk fire alarms, medium-risk fire alarms, and high-risk fire alarms; and by focusing on key time steps, it highlights the time points when fire and equipment abnormal events occur, providing support for timely alarm and response.

[0029] By performing the above operations, the traditional fire monitoring and early warning system is difficult to adapt to the dynamic fluctuations of complex building environments, resulting in low accuracy of fire signal detection, insufficient recognition of hidden fire precursors and equipment abnormal patterns, and thus difficulty in accurately warning of potential fire risks. This solution is based on multi-scale time series feature extraction and adaptive time step scaling mechanism to achieve synchronous detection of instantaneous fire signals and periodic equipment anomalies, and accurately locate the time point when a fire or abnormal event occurs; based on time domain feature focusing and nonlinear sensor intensity changes and dynamic trend adjustment mechanism, high-precision positioning of fire precursors and equipment failures can be achieved, and then small but potentially risky fire signals and equipment abnormal patterns can be identified, comprehensively improving the response effect and risk warning capability of the building fire monitoring system.

[0030] Example 4, see Figure 1 and Figure 2 , this embodiment is based on the above embodiment, and the laboratory safety management model establishment module also includes the following contents:

[0031] Anomaly detection unit; The anomaly detection unit captures abnormal changes in environmental monitoring data, equipment operation data, and spatial status data; specifically, it includes building a dual-channel joint feature extraction network, outputting sensor intensity changes and sensor dynamic trend prediction results respectively through a shared feature extraction layer; and reconstructing the time domain signal through ISTFT, calculating the reconstruction error to determine the abnormal fire signal and equipment failure behavior of the corresponding time step; introducing a secondary screening mechanism, for the abnormal points screened out initially, combining the local reconstruction error and historical fire statistical characteristics to calculate the comprehensive anomaly score, and making a final judgment on the abnormal points based on this; expressed as: ; ; ; ; ; ; expressed as: ; ;in, is the predicted intensity of sensor intensity change; is the predicted dynamic trend; SFN(·) is a dual-channel joint feature extraction network; is the reconstructed time domain signal; is the window function; L is the reconstruction window length, ranging from [10,50], and m is the window index; and is the predicted value under the corresponding window; R is the step size of the window movement, and its value range is [1,5]; Anomaly1 and Anomaly are the initial anomaly judgment and the final anomaly judgment results respectively; 1 is abnormal and 0 is normal; is the input original time domain signal; It is the signal reconstructed by the network of sensor intensity changes and sensor dynamic trends; is the dynamic error threshold; and are the historical mean and standard deviation of the reconstruction error, respectively; is the adjustment factor, the value range is [0.1,1.0]; and is the weight coefficient, the value range is [0.01,1.0]; is the anomaly score value; is the local reconstruction error; and They are the number of historical abnormal statistics and the total number of historical statistics; is the abnormal threshold, the value range is [0.5, 2.0]; and is the adjustment parameter; and They are the local changes of sensor intensity change and sensor dynamic trend respectively; by analyzing the intensity change and dynamic trend of sensor data at the same time, the hidden fire signals and equipment abnormal patterns can be captured; by reconstructing the error, high-precision anomaly detection can be achieved, and potential fire risks, equipment failures and hidden security threats in fire monitoring can be quickly identified;

[0032] Loss function design unit; build early detection reward loss , for early detection of fire and equipment failure, the system gives a lower penalty for early warning errors, encouraging early capture of abnormal signals and issuing alarms; expressed as: ; Construct discrete error loss , ensuring that the prediction results are more concentrated on the key time steps; expressed as: ; get the final loss , expressed as: ; ; ; Update the model parameters based on the gradient descent algorithm, and then establish a building fire assessment model; Among them, is the predicted time step of fire and equipment failure; is the actual time step of fire and equipment failure; N is the total number of time steps; , and They are the number of times the predicted fire time step is concentrated after the actual fire time step, the predicted fire time step is concentrated before the actual fire time step, and the predicted fire time step is concentrated in the actual fire time step and completely overlaps; is the root mean square error; and is the loss weight; is the standard deviation of the abnormal time step error; mean(·) is the average value; is the early detection reward loss for the j-th input.

[0033] By performing the above operations, this solution introduces a secondary screening mechanism and a dynamic threshold adjustment strategy to address the problems that traditional fire monitoring and early warning systems are difficult to adapt to the dynamic changes of environmental and equipment data in different time periods, have a high false alarm rate and are severely affected by noise, as well as insufficient timeliness in predicting fire and equipment anomalies, which in turn limits the reliability and stability of the overall monitoring system. This solution significantly improves the accuracy and reliability of fire signal and equipment anomaly detection results by constructing an early detection reward loss and reducing the penalty for early exposure of fire precursor signals or equipment failures, thereby meeting the early warning needs for fire and abnormal equipment failures, and thereby comprehensively improving the early warning accuracy and response efficiency of the building fire monitoring system.

[0034] Example 5, see Figure 1 This embodiment is based on the above embodiment. The building fire monitoring module is based on the established building fire assessment model, collects building fire data in real time, and monitors the building fire according to the fire assessment results output by the building fire assessment model; when the output fire assessment results are medium-risk fire alarms and high-risk fire alarms, early warning processing is performed on management personnel.

[0035] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0036] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

[0037] The present invention and its embodiments are described above, and such description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design a structure and embodiment similar to the technical solution without creativity, which should belong to the protection scope of the present invention.

Claims

1. Building fire monitoring system based on the Internet of Things, characterized by: The system includes a data acquisition module, a building fire assessment model establishment module and a building fire monitoring module; The data acquisition module collects fire monitoring data of historical buildings; The building fire assessment model establishment module realizes the synchronous detection of instantaneous fire signals and periodic equipment anomalies through multi-scale time series feature extraction, adaptive time step scaling mechanism, dual-channel joint feature extraction network and loss function optimization, thereby completing the establishment of the building fire assessment model; The building fire monitoring module detects and warns users of real-time data based on the established building fire assessment model; The building fire assessment model establishment module includes the following contents: a multi-scale time series feature extraction unit; the execution operations are: sensor data channel downsampling: extracting the time step features of sensor data through a 1D convolution kernel and reducing the number of channels, and introducing an adaptive time step scaling mechanism to dynamically adjust the time step to capture instantaneous high-frequency signals of fire events and periodic abnormal signals of equipment failures; time series pooling downsampling: using time pooling to reduce the number of time steps; time series upsampling: restoring the time step through inverse convolution to retain the global time series features; Multi-scale feature aggregation: Brother feature blocks from different downsampling and upsampling paths are directly aggregated; And introduce the time domain feature focusing and dynamic updating mechanism to focus on the time point when abnormal events occur, highlighting the key time steps of fire and equipment failure; expressed as: ; ; ; ; ; ; ;in, is the original input time series data, t is the time step; is a one-dimensional convolution operation, k is the kernel size; is the sigmoid activation function; It is a one-dimensional pooling operation; and They are channel downsampling and timing upsampling outputs respectively; is the inverse convolution operation; k1 is the number of channels; and They are the results of multi-scale feature aggregation and time domain feature focusing; are features from different channels; and They are average pooling and maximum pooling operations respectively; is the final weighted output; is the adaptive time step scaling parameter; is the dynamic weight coefficient; the GRU module is used to capture the dependencies between time steps; and are the upper and lower limits of the time step scaling factor; It is the softplus function; is the rate of change of the current time step; and is the original time series data corresponding to the time step; is the smoothing term.

2. The building fire monitoring system based on the Internet of Things according to claim 1 is characterized in that: The building fire assessment model establishment module specifically includes the following contents: Multi-scale temporal feature extraction unit; Anomaly detection unit; specifically, it includes constructing a dual-channel joint feature extraction network, outputting the sensor intensity change and sensor dynamic trend prediction results respectively through the shared feature extraction layer; reconstructing the time domain signal through ISTFT, calculating the reconstruction error to determine the abnormal fire signal and equipment failure behavior of the corresponding time step; introducing a secondary screening mechanism, for the abnormal points screened out initially, combining the local reconstruction error and historical fire statistical characteristics to calculate the comprehensive anomaly score, and making a final judgment on the abnormal points based on this; expressed as: ; ; ; ; ; ; expressed as: ; ;in, is the predicted intensity of sensor intensity change; is the predicted dynamic trend; SFN(·) is a dual-channel joint feature extraction network; is the reconstructed time domain signal; is the window function; L is the reconstruction window length, and m is the window index; and is the predicted value under the corresponding window; R is the step size of the window movement; Anomaly1 and Anomaly are the initial anomaly judgment and the final anomaly judgment results respectively; 1 is abnormal and 0 is normal; is the input original time domain signal; It is the signal reconstructed by the network of sensor intensity changes and sensor dynamic trends; is the dynamic error threshold; and are the historical mean and standard deviation of the reconstruction error, respectively; is the adjustment factor; and is the weight coefficient; is the anomaly score value; is the local reconstruction error; and They are the number of historical abnormal statistics and the total number of historical statistics; is the abnormal threshold; and is the adjustment parameter; and They are the local changes of sensor intensity change and sensor dynamic trend, respectively; Loss function design unit.

3. The building fire monitoring system based on the Internet of Things according to claim 2 is characterized in that: The loss function design unit is to construct an early detection reward loss ; expressed as: ; Construct discrete error loss , ensuring that the prediction results are more concentrated on the key time step; expressed as: ; get the final loss , expressed as: ; ; ; Update the model parameters based on the gradient descent algorithm, and then establish a building fire assessment model; Among them, is the predicted time step of fire and equipment failure occurrence; is the actual time step of fire and equipment failure; N is the total number of time steps; , and They are the number of times the predicted fire time step is concentrated after the actual fire time step, the predicted fire time step is concentrated before the actual fire time step, and the predicted fire time step is concentrated in the actual fire time step and completely overlaps; is the root mean square error; and is the loss weight; is the standard deviation of the abnormal time step error; mean(·) is the average value; is the early detection reward loss for the j-th input.

4. The building fire monitoring system based on the Internet of Things according to claim 3 is characterized in that: In the data acquisition module, the fire monitoring data of the historical buildings includes environmental monitoring data, equipment operation data, space status data, time and fire assessment results; the fire assessment results are used as data labels; and the collected data are converted and standardized.

5. The building fire monitoring system based on the Internet of Things according to claim 4 is characterized in that: The building fire monitoring module is based on the established building fire assessment model, collects building fire data in real time, and monitors building fire according to the fire assessment results output by the building fire assessment model.

Images