Viewing, listening, touching and smelling integrated testing method

Abstract

The invention discloses a vision, listening, touching and smelling integrated testing method, and belongs to the technical field of public fire safety. The invention relates to a vision, listening, touching and smelling integrated testing method. The method comprises the following steps: step 1, collecting and classifying fire-fighting equipment related data in a public area; 2, preprocessing the data; 3, establishing a security test model; 4, obtaining a test result; and 5, obtaining an integrated test result of fire safety. According to the invention, the problem that risk assessment is inaccurate due to the fact that analysis cannot be carried out from multiple angles when risk assessment is carried out on a fire in the prior art is solved, test results of visual sense, auditory sense, tactile sense and olfaction sense are integrated, integrated test is carried out, test analysis from multiple angles is realized, and the risk assessment accuracy is improved. The accuracy of fire alarm can be remarkably improved, the fire occurrence rate is greatly reduced, an intelligent early warning network is achieved, all-weather fire prevention is provided, and intelligent fire protection development is promoted.

Description

Technical Field

[0001] The present invention relates to the technical field of public fire safety, and in particular to a method for integrating sight, hearing, touch and smell. Background Art

[0002] With the development of my country's social economy, the expansion of urban scale, the production of new industries and the growth of national gross assets, large shopping malls and high-rise buildings have become important symbols of cities, people's production and living gas, oil and electricity consumption has gradually increased, and urban fire losses have become increasingly greater.

[0003] The Chinese patent with publication number CN108053594A discloses a fire warning method for a fixed-point regional unit: analyzing historical fire data and comprehensively considering potential fire hazards and fire risk resistance capabilities, determining fire risk factors; determining risk factor indexes; establishing a fire risk level model; determining the fire risk index of the fixed-point regional unit; determining the fire risk level of the fixed-point regional unit and issuing a warning. According to the established fire risk level model, the fire risk level of the fixed-point regional unit is determined and a fire risk warning is issued. All fire hazards are summarized into three categories: fire inspection, fire acceptance and remote monitoring, and these three factors are added to the fire warning model. The method is reasonably designed and effectively improves the practical application and accuracy of fire warning.

[0004] In actual use, the above patent uses fire inspections, fire acceptance inspections and remote monitoring to provide fire warnings. However, it cannot analyze fire risks from multiple angles, resulting in inaccurate risk assessments. Therefore, it does not meet existing needs. We have proposed an integrated testing method of sight, hearing, touch and smell. Summary of the invention

[0005] The purpose of the present invention is to provide an integrated testing method for sight, hearing, touch and smell. By integrating the test results of vision, hearing, touch and smell, an integrated test is performed, and test analysis is performed from multiple angles. This can significantly improve the accuracy of fire alarms, significantly reduce the incidence of fires, realize an intelligent early warning network, provide all-weather fire prevention, and promote the development of smart fire protection, thereby solving the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solution: a method for integrated testing of sight, hearing, touch and smell, comprising the following steps:

[0007] Step 1: Use the collection module to collect data related to firefighting equipment in public areas, and classify the collected data according to visual data, auditory data, tactile data, and olfactory data;

[0008] Step 2: Preprocessing the tactile data, olfactory data, visual data and auditory data using a data processing module;

[0009] Step 3: Establish a safety test model, and add the pre-processed visual data, auditory data, collected tactile data, and olfactory data to the safety test model;

[0010] Step 4: Test the input visual data, auditory data, collected tactile data, and olfactory data respectively to obtain test results;

[0011] Step 5: Integrate the test results of visual data, auditory data, collected tactile data, and olfactory data to obtain an integrated test result of fire safety.

[0012] Preferably, the collection module comprises:

[0013] The visual module is used to continuously capture video data in public areas using the integrated security and fire protection camera, and to continuously collect smoke data in public areas using the smoke sensor;

[0014] An auditory module is used to continuously collect data related to the operation of fire alarm controllers in public areas and to identify whether there is fire-related sound data in the auditory data of public places;

[0015] The tactile module is used to continuously collect the fire source temperature data collected by the infrared temperature detectors in the public area;

[0016] Olfactory module, used to continuously collect odor data in public areas using bionic gas sensors:

[0017] The classification module is used to classify the collected data according to visual data, auditory data, tactile data and olfactory data.

[0018] Preferably, the data processing module includes:

[0019] The visual data processing module is used to analyze the video data taken by the integrated security and firefighting camera, and extract the fire information from the image data that meets the information extraction requirements;

[0020] An auditory data processing module is used to extract relevant features from the sound data to obtain the fire sound features, and to obtain the fire vibration features from the vibration data;

[0021] The olfactory data processing module is used to classify and process the collected odor data from public areas, and analyze the characteristic differences of the olfactory data at different fire stages after processing;

[0022] The tactile data processing module is used to pre-process the temperature data of the public area and then conduct in-depth analysis to extract the temperature change trend and abnormal fluctuation data.

[0023] Preferably, the visual data processing module includes:

[0024] Analyze the video data captured by the integrated security and fire cameras to obtain unlabeled image data;

[0025] Filter low-quality images in unlabeled image data, and process the image data to obtain image data whose image quality meets the information extraction requirements;

[0026] The fire information in the image data is extracted, and the flames and smoke in the image data are labeled.

[0027] Preferably, the extracting of fire information from the image data specifically includes:

[0028] Image sampling and image segmentation are performed according to the frame difference method. The suspicious area image is obtained by subtracting adjacent frame images, and the flame and smoke features of the suspicious area are extracted.

[0029] The image information entropy and brightness change rate of the suspicious area are calculated to obtain the flame feature vector, and the area growth rate and centroid deviation of the suspicious area are calculated to obtain the smoke feature vector;

[0030] The extracted flame feature vector and smoke feature vector are used to obtain the fire information judgment result through the fixed threshold detection method.

[0031] Preferably, the auditory data processing module comprises:

[0032] Extract voiceprint features, sound frequency, volume changes, and keyword recognition results from the sound data to obtain a fire sound spectrum, and obtain sound features based on the fire sound spectrum;

[0033] Through the Fourier transform method, the vibration signal is converted from the time domain to the frequency domain, so as to obtain the main frequency components of the vibration signal, capture the different frequency components in the vibration signal, and record the intensity of each component;

[0034] At the same time, the start and end time points of the vibration signal are recorded, and the duration of the vibration event is calculated according to the start and end time points of the vibration signal;

[0035] The vibration characteristics of the fire are obtained based on the intensity of different frequency components in the vibration signal and the duration of the vibration event.

[0036] Preferably, the olfactory data processing module specifically includes:

[0037] The odor data collected from public areas are sorted and organized according to different collection locations and different possible fire source types;

[0038] After the classification is completed, the outliers in the odor data are removed, and it is determined whether there is any missing data in the odor data. If there is any missing data, the missing data is filled using the filling mean method;

[0039] After processing, the cleaned data is analyzed using machine learning algorithms to explore potential patterns in the data and analyze the characteristic differences of olfactory data at different fire stages.

[0040] Preferably, the tactile data processing module specifically includes:

[0041] Preprocess the received public area temperature data, including removing noise and outlier processing;

[0042] Conduct in-depth analysis on the preprocessed data to extract the changing trends and abnormal fluctuation data of the temperature in the public areas.

[0043] Preferably, the establishing of the safety test model specifically includes:

[0044] Collect historical fire videos, historical fire temperatures, sound spectra, and historical vibration characteristics, and obtain video frames;

[0045] Obtain smoke samples, sound samples and temperature sample data collected by smoke detectors, infrared temperature detectors and fire alarm controllers;

[0046] Generate pre-check values ​​using smoke samples and samples of historical fire temperature, historical sound spectrum, and historical vibration characteristics;

[0047] Take one frame every 10 frames of the video to generate a detection frame, and calculate the flame feature value of each pixel point in the detection frame;

[0048] Calculate the smoke feature value of each pixel in the detection frame, select the flame feature value to calculate the flame judgment dimension;

[0049] The smoke feature value is selected to calculate the smoke judgment dimension, and the flame judgment dimension, smoke judgment dimension, fire temperature, sound spectrum and vibration feature samples are respectively input into the classifier for training to generate a fire test model.

[0050] Preferably, the safety test model specifically includes:

[0051] The visual data testing module is used to identify smoke and smoke volume of fire information in the image data, and to locate open fire and analyze fire intensity of fire information in the image data to obtain visual test results;

[0052] The auditory data testing module is used to combine the sound characteristics and the vibration characteristics to test whether there is an abnormal fire sound in the current public area according to the sound characteristics and the vibration characteristics;

[0053] The tactile data test module is used to convert the temperature change trend of the public area and a frame of temperature data with abnormal fluctuations into a matrix, correspond to the actual picture and calculate the average temperature of the frame, and compare the average with the current temperature average to obtain the tactile test result;

[0054] The olfactory data testing module is used to train known fire olfactory data and non-fire olfactory data according to the characteristic differences of olfactory data at different fire stages, test whether there is fire data in the olfactory data, and judge the size of the fire.

[0055] Compared with the prior art, the present invention has the following beneficial effects:

[0056] The present invention can significantly improve the accuracy of fire alarms by integrating the test results of vision, hearing, touch and smell to realize test analysis from multiple angles. When the smoke sensing device alarms, the integrated security and firefighting camera can be immediately linked to view the surrounding environment of the alarm point, thereby improving the accuracy of the smoke sensing device alarm. An intelligent early warning network can be formed to greatly reduce the incidence of fires, and 24-hour uninterrupted fire prevention can be provided. The alarm can be sounded even when there is no one on duty at night, which promotes the development of smart fire protection, enables flames to be discovered and extinguished in the budding stage, and the public safety system has also achieved a qualitative improvement. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a schematic diagram of the integrated testing method of sight, hearing, touch and smell of the present invention;

[0058] Figure 2 It is a schematic diagram of the module of the integrated testing method of sight, hearing, touch and smell of the present invention. DETAILED DESCRIPTION

[0059] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] In order to solve the problem that the existing patents cannot analyze fire risk assessment from multiple angles during fire inspection, fire inspection and remote monitoring in actual use, resulting in inaccurate risk assessment, please refer to Figure 1-Figure 2, this embodiment provides the following technical solutions:

[0061] The integrated test method of sight, hearing, touch and smell includes the following steps:

[0062] Step 1: Use the collection module to collect data related to firefighting equipment in public areas, and classify the collected data according to visual data, auditory data, tactile data and olfactory data. Public areas include parks, apartments and shopping malls. Firefighting equipment includes integrated security and firefighting cameras, smoke detectors, infrared temperature detectors and fire alarm controllers. The integrated security and firefighting cameras are combined with the infrared sensing functions of video and infrared temperature detectors. When the smoke sensing equipment alarms, the integrated security and firefighting cameras can be immediately linked to view the surrounding environment of the alarm point, thereby improving the accuracy of the smoke sensing equipment alarm;

[0063] Step 2: Preprocessing the tactile data, olfactory data, visual data and auditory data using a data processing module;

[0064] Step 3: Establish a safety test model, and add the pre-processed visual data, auditory data, collected tactile data, and olfactory data to the safety test model;

[0065] Step 4: The safety test model tests the input visual data, auditory data, collected tactile data and olfactory data respectively to obtain the test results of the visual data, auditory data, collected tactile data and olfactory data. By testing the visual data, auditory data, collected tactile data and olfactory data, it is possible to test and analyze public fire safety from multiple angles, thereby improving the accuracy of the test and analysis results, and thus significantly improving the accuracy of fire alarms and greatly reducing the incidence of fires.

[0066] Step 5: Integrate the test results of visual data, auditory data, collected tactile data and olfactory data to obtain an integrated test result of fire safety. By integrating the test results of vision, hearing, touch and smell, the accuracy of fire alarms can be significantly improved.

[0067] Collection modules, including:

[0068] The visual module is used to continuously capture video data in public areas using the integrated security and fire protection camera, and to continuously collect smoke data in public areas using the smoke sensor;

[0069] An auditory module is used to continuously collect data related to the operation of fire alarm controllers in public areas and to identify whether there is fire-related sound data in the auditory data of public places;

[0070] The tactile module is used to continuously collect the fire source temperature data collected by the infrared temperature detectors in the public area;

[0071] Olfactory module, used to continuously collect odor data in public areas using bionic gas sensors:

[0072] The classification module is used to classify the collected data according to visual data, auditory data, tactile data and olfactory data.

[0073] Data processing module, including:

[0074] The visual data processing module is used to analyze the video data taken by the integrated security and firefighting camera, and extract the fire information from the image data that meets the information extraction requirements;

[0075] The auditory data processing module is used to extract relevant features from the sound data to obtain the fire sound features, and obtain the fire vibration features according to the intensity of different frequency components in the vibration signal and the duration of the vibration event;

[0076] The olfactory data processing module is used to classify and process the odor data collected from public areas.

[0077] After processing, the characteristic differences of olfactory data in different fire stages are analyzed, such as smoldering stage, open flame stage, etc.

[0078] The tactile data processing module is used to pre-process the received public area temperature data and then conduct in-depth analysis to extract the change trend and abnormal fluctuation data of the public area temperature.

[0079] Visual data processing module, including:

[0080] Parse the video data captured by the integrated security and firefighting camera to obtain unlabeled image data, where the unlabeled image data includes structured information, and the video stream data includes video data obtained by real-time monitoring of the network camera in the video surveillance scene;

[0081] Filter low-quality images in unlabeled image data, and perform deduplication, random cropping, mosaic, noise, color conversion, brightness adjustment, and geometric transformation on the image data to obtain image data with image quality that meets the requirements of information extraction;

[0082] Extract fire information from image data, including suspected fire areas and fire features, extract the color, shape and movement trajectory of the flame, and mark the flame and smoke in the image data.

[0083] Extract fire information from image data, including:

[0084] Image sampling and image segmentation are performed according to the frame difference method. The suspicious area image is obtained by subtracting adjacent frame images, and the flame and smoke features of the suspicious area are extracted.

[0085] The image information entropy and brightness change rate of the suspicious area are calculated to obtain the flame feature vector, and the area growth rate and centroid deviation of the suspicious area are calculated to obtain the smoke feature vector;

[0086] The extracted flame feature vector and smoke feature vector are used to obtain the fire information judgment result through the fixed threshold detection method.

[0087] Auditory data processing module, including:

[0088] Extract voiceprint features, sound frequency, volume changes, and keyword recognition results from the sound data to obtain a fire sound spectrum, and obtain sound features based on the fire sound spectrum;

[0089] Through the Fourier transform method, the vibration signal is converted from the time domain to the frequency domain, so as to obtain the main frequency components of the vibration signal, capture the different frequency components in the vibration signal, and record the intensity of each component;

[0090] At the same time, the start and end time points of the vibration signal are recorded, and the duration of the vibration event is calculated according to the start and end time points of the vibration signal;

[0091] The vibration characteristics of the fire are obtained based on the intensity of different frequency components in the vibration signal and the duration of the vibration event.

[0092] Olfactory data processing module, specifically including:

[0093] The odor data collected from public areas are classified and sorted according to different collection locations and different possible fire source types to facilitate subsequent analysis and processing.

[0094] After the classification is completed, the outliers in the odor data are removed, and it is determined whether there is any missing data in the odor data. If there is any missing data, the missing data is filled using the filling mean method;

[0095] After processing, the cleaned data is analyzed using machine learning algorithms to explore the potential patterns in the data, analyze the characteristic differences of olfactory data in different fire stages, and improve the ability to identify fires.

[0096] The tactile data processing module specifically includes:

[0097] The received public area temperature data is preprocessed, including noise removal and outlier processing to improve the reliability of the data.

[0098] Conduct in-depth analysis on the preprocessed data to extract the changing trends and abnormal fluctuation data of the temperature in the public areas.

[0099] Establish a security testing model, including:

[0100] Collect historical fire videos, historical fire temperature, sound spectrum, and historical vibration characteristics, and obtain video frames with a frame rate of 25 frames per second. Select 1,000 flame samples from the video frames.

[0101] Obtain smoke samples, sound samples and temperature sample data collected by smoke detectors, infrared temperature detectors and fire alarm controllers;

[0102] Generate pre-check values ​​using 1000 smoke samples and 4000 samples of historical fire temperature, historical sound spectrum and historical vibration characteristics;

[0103] Take one frame every 10 frames of the video to generate a detection frame, and calculate the flame feature value of each pixel point in the detection frame;

[0104] The smoke generated smoke feature value of each pixel in the detection frame is calculated, and 10 flame feature values ​​are selected to calculate 55 flame judgment dimensions;

[0105] Ten smoke feature values ​​were selected to calculate 55 smoke judgment dimensions. The flame judgment dimension, smoke judgment dimension, fire temperature, sound spectrum and vibration feature samples were respectively input into the classifier for training to generate a fire test model.

[0106] Security testing model, including:

[0107] The visual data testing module is used to identify smoke and smoke volume of fire information in image data, and to locate open flames and analyze fire intensity of fire information in image data, to obtain the smoke area and the dynamic area and open flame area of ​​dense fog area that change with time, to perform feature fitting on the smoke area, the dynamic area of ​​dense fog area that changes with time, the average width of the dense smoke area that changes dynamically with time, the axis length of the optical flow direction, and the open flame area, to obtain visual test results.

[0108] The auditory data test module is used to combine the sound characteristics and vibration characteristics to test whether there is an abnormal fire sound in the current public area based on the sound characteristics and vibration characteristics.

[0109] The tactile data test module is used to convert the temperature change trend of the public area and a frame of temperature data with abnormal fluctuations into a matrix, correspond to the actual picture and calculate the average temperature of the frame, and compare the average temperature with the current average temperature to obtain the tactile test result.

[0110] Specifically, the temperature change trend of the public area and the abnormal fluctuation of a frame of temperature data are converted into a matrix, corresponding to the actual picture and calculating the average temperature of the frame, and comparing the average with the current temperature average;

[0111] If the temperature difference does not exceed 1 degree, the average temperature of the frame and the average temperature of the previous frame are averaged to obtain the new current temperature average. The pixel temperature in the temperature matrix is ​​represented as brightness information;

[0112] From pure black to pure white, there are 256 levels corresponding to -56℃ to 200℃, which depicts the temperature image. The pixels in the temperature matrix that exceed the set temperature value are assigned 1, and the remaining pixels are assigned 0;

[0113] After converting into a binary image, the region and area of ​​the overheated image can be obtained by calculating the function, and the tactile test result can be obtained.

[0114] The olfactory data testing module is used to train known fire olfactory data and non-fire olfactory data according to the characteristic differences of olfactory data at different fire stages, test whether there is fire data in the olfactory data, and judge the size of the fire.

[0115] In summary, the integrated sight, hearing, touch and smell testing method of the present invention can significantly improve the accuracy of fire alarms by integrating the test results of vision, hearing, touch and smell. The integrated safety and firefighting camera combines the infrared sensing function of video and infrared temperature detectors, and can immediately link the integrated safety and firefighting camera to view the surrounding environment of the alarm point when the smoke sensing equipment alarms, thereby improving the accuracy of the smoke sensing equipment alarm, forming an intelligent early warning network, greatly reducing the incidence of fire, and can detect fire points and issue alarms in time, so that surrounding rescue forces can be dispatched to the scene to extinguish the fire in the early stage of the fire. It can provide 24-hour uninterrupted fire prevention and can issue alarms even when there is no one on duty at night, which promotes the development of smart firefighting, so that flames can be discovered and extinguished in the budding stage, making the security and firefighting industries develop faster and more intelligently, and the public safety system has also achieved a qualitative improvement.

[0116] 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.

[0117] 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.

Claims

1. A visual, auditory, tactile and olfactory integrated testing method, characterized in that: The following steps are involved: Step 1: Use the collection module to collect fire equipment data in public areas, and classify the collected data according to visual data, auditory data, tactile data, and olfactory data; Step 2: Preprocessing the tactile data, olfactory data, visual data and auditory data using a data processing module; Step 3: Establish a safety test model, and add the pre-processed visual data, auditory data, collected tactile data, and olfactory data to the safety test model; Step 4: The input visual data, auditory data, collected tactile data and olfactory data are tested respectively to obtain test results; Step 5: Integrate the test results of visual data, auditory data, collected tactile data, and olfactory data to obtain an integrated test result of fire safety.

2. The integrated testing method of sight, hearing, touch and smell according to claim 1, characterized in that: The collection module comprises: The visual module is used to continuously capture video data in public areas using the integrated security and fire protection camera, and to continuously collect smoke data in public areas using the smoke sensor; The auditory module is used to continuously collect the working data of the fire alarm controller in the public area and to identify whether there is fire sound data in the auditory data of the public place; The tactile module is used to continuously collect the fire source temperature data collected by the infrared temperature detectors in the public area; Olfactory module, used to continuously collect odor data in public areas using bionic gas sensors: The classification module is used to classify the collected data according to visual data, auditory data, tactile data and olfactory data.

3. The integrated testing method of sight, hearing, touch and smell according to claim 1, characterized in that: The data processing module comprises: The visual data processing module is used to analyze the video data captured by the integrated security and firefighting camera, and extract the fire information from the image data that meets the information extraction requirements; An auditory data processing module is used to extract relevant features from the sound data to obtain the fire sound features, and to obtain the fire vibration features from the vibration data; The olfactory data processing module is used to classify and process the collected odor data from public areas, and analyze the characteristic differences of the olfactory data at different fire stages after processing; The tactile data processing module is used to pre-process the temperature data of the public area and then conduct in-depth analysis to extract the temperature change trend and abnormal fluctuation data.

4. The integrated testing method of sight, hearing, touch and smell according to claim 3, characterized in that: The visual data processing module comprises: Analyze the video data captured by the integrated security and fire cameras to obtain unlabeled image data; Filter low-quality images in unlabeled image data, and process the image data to obtain image data whose image quality meets the information extraction requirements; The fire information in the image data is extracted, and the flames and smoke in the image data are labeled.

5. The integrated testing method of sight, hearing, touch and smell according to claim 4, characterized in that: The extracting of fire information from the image data specifically includes: Image sampling and image segmentation are performed according to the frame difference method. The suspicious area image is obtained by subtracting adjacent frame images, and the flame and smoke features of the suspicious area are extracted. The image information entropy and brightness change rate of the suspicious area are calculated to obtain the flame feature vector, and the area growth rate and centroid deviation of the suspicious area are calculated to obtain the smoke feature vector; The extracted flame feature vector and smoke feature vector are used to obtain the fire information judgment result through the fixed threshold detection method.

6. The integrated testing method of sight, hearing, touch and smell according to claim 3, characterized in that: The auditory data processing module comprises: Extract voiceprint features, sound frequency, volume changes, and keyword recognition results from the sound data to obtain a fire sound spectrum, and obtain sound features based on the fire sound spectrum; Through the Fourier transform method, the vibration signal is converted from the time domain to the frequency domain, so as to obtain the main frequency components of the vibration signal, capture the different frequency components in the vibration signal, and record the intensity of each component; At the same time, the start and end time points of the vibration signal are recorded, and the duration of the vibration event is calculated according to the start and end time points of the vibration signal; The vibration characteristics of the fire are obtained based on the intensity of different frequency components in the vibration signal and the duration of the vibration event.

7. The integrated testing method of sight, hearing, touch and smell according to claim 3, characterized in that: The olfactory data processing module specifically includes: Classify and sort the odor data collected in public areas according to different collection locations and different possible fire source types; After the classification is completed, the outliers in the odor data are removed, and it is determined whether there is any missing data in the odor data. If there is any missing data, the missing data is filled using the filling mean method; After processing, the cleaned data is analyzed using machine learning algorithms to explore potential patterns in the data and analyze the characteristic differences of olfactory data at different fire stages.

8. The integrated testing method of sight, hearing, touch and smell according to claim 3, characterized in that: The tactile data processing module specifically includes: The received public area temperature data is preprocessed, and the preprocessing includes noise removal and outlier processing. Conduct in-depth analysis on the preprocessed data to extract the changing trends and abnormal fluctuation data of the temperature in the public areas.

9. The integrated testing method of sight, hearing, touch and smell according to claim 1, characterized in that: The establishment of a security test model specifically includes: Collect historical fire videos, historical fire temperatures, sound spectra, and historical vibration characteristics, and obtain video frames; Obtain smoke samples, sound samples and temperature sample data collected by smoke detectors, infrared temperature detectors and fire alarm controllers; Generate pre-check values ​​using smoke samples and samples of historical fire temperature, historical sound spectrum, and historical vibration characteristics; Take one frame every 10 frames of the video to generate a detection frame, and calculate the flame feature value of each pixel point in the detection frame; Calculate the smoke feature value of each pixel in the detection frame, select the flame feature value to calculate the flame judgment dimension; The smoke feature value is selected to calculate the smoke judgment dimension, and the flame judgment dimension, smoke judgment dimension, fire temperature, sound spectrum and vibration feature samples are respectively input into the classifier for training to generate a fire test model.

10. The integrated testing method of sight, hearing, touch and smell according to claim 5, characterized in that: The safety test model specifically includes: The visual data testing module is used to identify smoke and smoke volume of fire information in the image data, and to locate open fire and analyze fire intensity of fire information in the image data to obtain visual test results; The auditory data testing module is used to combine the sound characteristics and the vibration characteristics to test whether there is an abnormal fire sound in the current public area according to the sound characteristics and the vibration characteristics; The tactile data test module is used to convert the temperature change trend of the public area and a frame of temperature data with abnormal fluctuations into a matrix, correspond to the actual picture and calculate the average temperature of the frame, and compare the average with the current temperature average to obtain the tactile test result; The olfactory data testing module is used to train known fire olfactory data and non-fire olfactory data according to the characteristic differences of olfactory data at different fire stages, test whether there is fire data in the olfactory data, and judge the size of the fire.

Images