Fire early warning method and device based on intelligent model

Abstract

The embodiment of the invention provides a fire early warning method and device based on an intelligent model. The method comprises the following steps: acquiring fire historical data; wherein the fire historical data comprises fire historical image data, fire historical sound data, fire historical temperature data and water vapor evaporation degree data; extracting classification features of the historical fire data, and screening feature mutation data from the classification features; inputting the feature mutation data and the corresponding pre-divided fire early warning types into the first fire early warning training model to obtain a trained first fire early warning training model; inputting the first fire data into the trained first fire early warning training model to obtain output specific fire early warning type data; according to the method, different fire hazard types and sudden change characteristics among different fire behaviors are trained to obtain characteristic data among different fire behaviors and fire hazard types, so that the accuracy, timeliness and sensitivity of fire behavior identification are improved, and the diversity and accuracy of fire hazard early warning are improved.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a fire warning method based on an intelligent model, a fire warning device based on an intelligent model, a computer device and a computer-readable storage medium. Background Art

[0002] The method of warning fires through image recognition mainly relies on AI cameras, deep learning algorithms and real-time monitoring systems.

[0003] The development of modern science and technology has made the fire warning method based on image recognition more and more intelligent and efficient. In terms of specific implementation, AI cameras can be used for real-time monitoring, and deep learning algorithms can be used to detect smoke and flames. For example, AI cameras can be selected as the hardware basis to capture real-time video streams and perform preliminary analysis on video data. High-performance computing devices, such as GPU servers, are used to process large amounts of image data and run complex deep learning algorithms. The collected images are pre-processed, such as enhancement and noise reduction, to improve the effect of subsequent analysis and recognition. Image grayscale conversion and histogram equalization are used to improve image quality and contrast, making the features more obvious. Feature extraction and classification: Extract the features of smoke and flames from the processed images, including color, shape, area, and motion trajectory. Use machine learning algorithms (such as support vector machines, SVM) or deep learning algorithms (such as convolutional neural networks, CNN) to classify and identify features. Combined with AI technology, real-time monitoring and early warning of fires can be achieved, and fire conditions can be discovered in a timely manner and corresponding measures can be taken. However, the existing fire warning function is still not accurate enough and has low accuracy. Summary of the invention

[0004] In view of the above problems, embodiments of the present invention are proposed to provide a fire warning method based on an intelligent model, a fire warning device based on an intelligent model, a computer device and a computer-readable storage medium that overcome the above problems or at least partially solve the above problems.

[0005] In order to solve the above problems, an embodiment of the present invention discloses a fire warning method based on an intelligent model, the method comprising:

[0006] Obtain fire history data;

[0007] The fire history data includes fire history image data, fire history sound data, fire history temperature data, and water vapor evaporation degree data;

[0008] Extracting classification features of the fire history data, and filtering out feature mutation data from the classification features;

[0009] Inputting the characteristic mutation data and the corresponding pre-classified fire warning types into the first fire warning training model to obtain a trained first fire warning training model;

[0010] The first fire data is input into the trained first fire warning training model to obtain output specific fire warning type data.

[0011] Preferably, the method comprises:

[0012] Inputting the characteristic mutation data and the corresponding pre-divided fire diffusion probability into the second fire warning training model to obtain a trained second fire warning training model;

[0013] The second fire data is input into the trained second fire warning training model to obtain output specific fire spread probability data.

[0014] Preferably, the method comprises:

[0015] Inputting the characteristic mutation data and the corresponding pre-divided fire hazard data into the third fire warning training model to obtain a trained third fire warning training model;

[0016] The third fire data is input into the trained third fire warning training model to obtain output specific fire hazard data.

[0017] Preferably, extracting classification features of the fire history data and screening out feature mutation data from the classification features includes:

[0018] Extracting image classification data of the fire history image data;

[0019] and / or, extracting sound classification data of the fire history sound data;

[0020] and / or, extracting temperature classification data of the fire history temperature data;

[0021] And / or, extracting water vapor evaporation classification data from the water vapor evaporation degree data.

[0022] Preferably, extracting classification features of the fire history data and screening out feature mutation data from the classification features includes:

[0023] Extracting image mutation classification features from the image classification data;

[0024] and / or, extracting sound mutation classification features from the sound classification data;

[0025] and / or, extracting temperature mutation classification features from the temperature classification data;

[0026] And / or, extracting water vapor evaporation mutation classification features from the water vapor evaporation classification data.

[0027] Preferably, the step of inputting the characteristic mutation data and the corresponding pre-classified fire warning types into the first fire warning training model to obtain the trained first fire warning training model includes:

[0028] Converting the characteristic mutation data and the corresponding pre-divided fire warning types into a warning mutation characteristic vector;

[0029] Converting the early warning mutation feature vector into first mutation dimension data;

[0030] The mutation dimension data is input into the convolution layer, pooling layer, fully connected layer and activation layer of the first fire warning training model to obtain the trained first fire warning training model.

[0031] Preferably, the second fire warning training model includes a Bayesian neural network; the step of inputting the characteristic mutation data and the corresponding pre-divided fire spread probability into the second fire warning training model to obtain the trained second fire warning training model includes:

[0032] Converting the characteristic mutation data and the corresponding pre-divided fire diffusion probability into a diffusion mutation characteristic vector;

[0033] Converting the diffusion mutation feature vector into second mutation dimension data;

[0034] The mutation dimension data is input into the input layer, hidden layer, activation layer, and output layer of the Bayesian neural network to obtain a trained Bayesian neural network.

[0035] The embodiment of the present invention discloses a fire warning device based on an intelligent model, the device comprising:

[0036] A fire history data acquisition module, used to acquire fire history data; wherein the fire history data includes fire history image data, fire history sound data, fire history temperature data, and water vapor evaporation degree data;

[0037] An extraction module, used to extract classification features of the fire history data, and filter out feature mutation data from the classification features;

[0038] A first training module, used for inputting the characteristic mutation data and the corresponding pre-classified fire warning types into the first fire warning training model to obtain a trained first fire warning training model;

[0039] The first output module is used to input the first fire data into the trained first fire warning training model to obtain output specific fire warning type data.

[0040] An embodiment of the present invention discloses a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above-mentioned fire warning method based on the intelligent model when executing the computer program.

[0041] The embodiment of the present invention discloses a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned fire warning method based on the intelligent model are implemented.

[0042] The embodiments of the present invention include the following advantages:

[0043] In an embodiment of the present invention, fire history data is obtained; wherein the fire history data includes fire history image data, fire history sound data, fire history temperature data, and water vapor evaporation degree data; classification features of the fire history data are extracted, and feature mutation data are screened out from the classification features; the feature mutation data and the corresponding pre-divided fire warning types are input into the first fire warning training model to obtain a trained first fire warning training model; the first fire data is input into the trained first fire warning training model to obtain output specific fire warning type data; the mutation features between different fire types and different fire intensities are trained to obtain feature data between different fire intensities and fire types, thereby improving the accuracy, timeliness and sensitivity of fire identification, and improving the diversity and accuracy of fire warnings. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0045] Figure 1 is a schematic diagram of an embodiment of a fire early warning method based on an intelligent model according to an embodiment of the present invention;

[0046] Figure 2 is a schematic diagram of a fire warning device based on an intelligent model according to an embodiment of the present invention;

[0047] Figure 3 The diagram is a diagram of the internal structure of a computer device according to an embodiment. DETAILED DESCRIPTION

[0048] In order to make the technical problems, technical solutions and beneficial effects solved by the embodiments of the present invention more clearly understood, the embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0049] Reference Figure 1 , shows a schematic diagram of an embodiment of a fire warning method based on an intelligent model according to an embodiment of the present invention, which may specifically include the following steps:

[0050] Step S101, obtaining fire history data; wherein the fire history data includes fire history image data, fire history sound data, fire history temperature data, and water vapor evaporation degree data;

[0051] In an embodiment of the present invention, the method can be applied to run on a terminal, which can be various computers, laptops, etc. The embodiment of the present invention does not limit the specific type of the terminal. The operating system of the terminal may include Android, Harmony OS, IOS, Windows Phone, Windows, etc., and the present invention does not impose too many restrictions on this.

[0052] In an embodiment of the present invention, the terminal may first acquire fire history data; specifically, the fire history data may include fire history image data, fire history sound data, fire history temperature data, water vapor evaporation degree data, etc., and may also include other types of fire history data, and the embodiment of the present invention does not impose excessive restrictions on this.

[0053] In addition, the types of fire may include solid material fires, liquid or fusible solid material fires, gas fires, metal fires, electrical fires, and fires of cooking materials in cooking utensils. Of course, other types of fires may also be included, and the embodiments of the present invention do not impose excessive restrictions on this.

[0054] Specifically, the fire history image data refers to the image data of different stages of different types of fires; the fire history sound data refers to the sound data of different stages of different types of fires; the fire history temperature data refers to the temperature data of different stages of different types of fires; and the water vapor evaporation degree data refers to the water vapor evaporation data of different stages of different types of fires.

[0055] For example, the fire history image data may be image data at the beginning of a solid material fire, first image data during the burning of a solid material fire, second image data during the burning of a solid material fire, third image data during the burning of a solid material fire, and image data at the end of a solid material fire; it may be collected by an image collector such as a camera.

[0056] Furthermore, the fire history sound data may be the sound data at the beginning of a gas fire, the first sound data during the combustion of a gas fire, the second sound data during the combustion of a gas fire, the third sound data during the combustion of a gas fire, and the sound data at the end of a gas fire, which may be audio data obtained by a recording device.

[0057] Specifically, the fire history temperature data can be the temperature data at the beginning of the gas fire, the first temperature data during the combustion of the gas fire, the second temperature data during the combustion of the gas fire, the third temperature data during the combustion of the gas fire, and the temperature data at the end of the gas fire, which can be data recorded by a temperature sensor.

[0058] Specifically, the fire history water vapor evaporation data can be the moisture content data of the air at the beginning of the metal fire, the first moisture content data of the air during the burning of the metal fire, the second moisture content data of the air during the burning of the metal fire, the third moisture content data of the air during the burning of the metal fire, and the moisture content data of the air at the end of the metal fire, which can be the data recorded by the air humidity sensor.

[0059] The above data types are only several examples of the embodiments of the present invention, and may also include other fire-related environment or other data. The embodiments of the present invention do not impose excessive restrictions on the types and quantities of data.

[0060] Step S102, extracting classification features of the fire history data, and filtering out feature mutation data from the classification features;

[0061] Furthermore, after obtaining various fire history data, classification features of various fire history data can be further extracted, and feature mutation data can be screened out from the classification features. Specifically, the extraction of classification features of the fire history data and the screening out of feature mutation data from the classification features include: extracting image classification data of the fire history image data; and / or, extracting sound classification data of the fire history sound data; and / or, extracting temperature classification data of the fire history temperature data; and / or, extracting water vapor evaporation classification data of the water vapor evaporation degree data.

[0062] In an embodiment of the present invention, the classification feature may include image classification data, sound classification data, temperature classification data and water vapor evaporation classification data. The image classification data may be flame feature data of different stages in the fire history image data of various types of fires, such as flame features in the first image data of solid material fire combustion, the second image data of solid material fire combustion, and the third image data of solid material fire combustion. The flame features may include outer flame features, middle flame features and flame core features.

[0063] The sound classification data may refer to the sound feature data at different stages of various types of fires, and the sound classification data is associated with different combustion stages of the fire burning objects; for example, the sound feature at the beginning of a fire is the sound of a small burning sound of the burning object, and the sound feature when the fire is large is the sound of a deep burning sound of the burning object, which may be accompanied by the sound of an object bursting and collapsing; specifically, the first sound data in a gas fire burning, the second sound data in a gas fire burning, the third sound data in a gas fire burning, etc.;

[0064] Temperature classification data may refer to temperature variation data at different stages in various types of fires, and the temperature characteristic classification data is associated with different combustion stages of the fire combustion materials; for example, the temperature is lower at the beginning of a fire and higher when the fire is larger, and a temperature chart can be generated based on the temperature characteristic classification data at different stages of the fire. In a preferred embodiment of the present invention, water vapor evaporation classification data may be water vapor evaporation data at different stages in various types of fires, such as water vapor evaporation characteristics of the first moisture content data of the air during the combustion of a metal fire, the second moisture content data of the air during the combustion of a metal fire, and the third moisture content data of the air during the combustion of a metal fire; in the embodiment of the present invention, relevant characteristics of water vapor evaporation degree data are introduced to improve the prediction accuracy and diversity of fire data and improve the accuracy of prediction results.

[0065] In a preferred embodiment of the embodiments of the present invention, the extracting of classification features of the fire history data and filtering out feature mutation data from the classification features include: extracting image mutation classification features from the image classification data; and / or, extracting sound mutation classification features from the sound classification data; and / or, extracting temperature mutation classification features from the temperature classification data; and / or, extracting water vapor evaporation mutation classification features from the water vapor evaporation classification data.

[0066] Further applied to the embodiments of the present invention, after obtaining the classification features of the fire history data, feature mutation data can be screened out from the classification features. Specifically, the feature mutation data can include image mutation classification features, sound mutation classification features, temperature mutation classification features, and water vapor evaporation mutation classification features, etc. The embodiments of the present invention do not impose too many restrictions on this. It should be noted that the above-mentioned classification features, such as image classification data, sound classification data, temperature classification data, and water vapor evaporation classification data, etc., are respectively associated with different stages of the fire, that is, the feature mutation data screened out from the classification features are also associated with different stages of the fire.

[0067] Specifically, multiple thresholds of different types can be preset to filter out feature mutation data; for example, the preset light intensity threshold can be a change of more than 300lx, and the light intensity at different stages in the fire history image data of various types of fires is ranked as follows: a lx, b lx, c lx, d lx, e lx, flx, and the change between each two data is compared. If the change between b lx and c lx exceeds 300lx, that is, exceeds the preset light intensity threshold, then c lx can be determined as the corresponding image mutation classification feature;

[0068] Alternatively, the preset temperature threshold can be set as a change of more than 20 degrees Celsius, and the temperature change data at different stages of each type of fire are sorted as follows: A degrees Celsius, B degrees Celsius, C degrees Celsius, D degrees Celsius, E degrees Celsius, and F degrees Celsius. The change between each two data is compared. For example, if the change between C degrees Celsius and D degrees Celsius exceeds 20 degrees Celsius, that is, exceeds the preset temperature threshold, then D degrees Celsius can be determined as the corresponding temperature mutation classification feature.

[0069] It should be noted that the above-mentioned classification features and feature mutation data are only a few examples of the embodiments of the present invention, and may also include other different types of classification features and feature mutation data associated with fire types and fire stages. The embodiments of the present invention do not impose too many restrictions on this.

[0070] Step S103: inputting the characteristic mutation data and the corresponding pre-classified fire warning types into the first fire warning training model to obtain a trained first fire warning training model;

[0071] In the embodiment of the present invention, after the characteristic mutation data is obtained, the sample data is formed with the corresponding pre-divided fire warning types; the sample data is applied to the first fire warning training model to obtain the trained first fire warning training model;

[0072] It should be noted that the fire warning types may refer to warning alarm types divided into different levels according to the duration of the fire, the scale of the fire or the area where the fire spreads.

[0073] For example, fire warning types can be divided into emergency large-scale fires, large-scale fires, slow small-scale fires, small-scale fires, etc., and can also be divided into other fire warning types. The embodiments of the present invention do not impose too many restrictions on this. Specifically, the feature mutation data and the corresponding pre-divided fire warning types are input into the first fire warning training model to obtain the trained first fire warning training model, including: converting the feature mutation data and the corresponding pre-divided fire warning types into warning mutation feature vectors; converting the warning mutation feature vectors into first mutation dimension data; inputting the mutation dimension data into the convolution layer, pooling layer, fully connected layer, and activation layer of the first fire warning training model to obtain the trained first fire warning training model.

[0074] In the embodiment of the present invention, the first fire warning training model may be a convolutional neural network model, and of course, may also be other types of supervised neural network models, and the embodiment of the present invention does not impose too many restrictions on this; specifically, the first fire warning training model may include a convolutional layer, a pooling layer, a fully connected layer, an activation layer, etc., and the embodiment of the present invention does not impose too many restrictions on the specific structure of the first fire warning training model;

[0075] In addition, the method of converting the warning mutation feature vector into the first mutation dimension data can be a method of performing data conversion through a convolution kernel. The embodiment of the present invention does not impose too many restrictions on this. The above-mentioned mutation feature screening method can reduce overfitting, improve the accuracy and stability of the machine learning model, and improve the generalization ability.

[0076] Step S104: input the first fire data into the trained first fire warning training model to obtain output specific fire warning type data.

[0077] In actual application in the embodiment of the present invention, after obtaining the trained first fire warning training model, the first fire data can be input into the trained first fire warning training model to obtain output specific fire warning type data; the first fire data can be data such as images, sounds, temperatures or air water vapor content of an area containing various flammable objects. The embodiment of the present invention does not impose too many restrictions on this. In the embodiment of the present invention, specific fire warning type data can be output to a display device, such as outputting one of the fire warning types (emergency large-scale fire, large-scale fire, slow small-scale fire, small-scale fire) to the screen, and issuing an alarm message to prompt the user.

[0078] In a preferred embodiment of the embodiments of the present invention, the method includes: inputting the characteristic mutation data and the corresponding pre-divided fire spread probability into the second fire warning training model to obtain a trained second fire warning training model; inputting the second fire data into the trained second fire warning training model to obtain output specific fire spread probability data.

[0079] The pre-divided fire spread probability refers to the probability of fire spreading and spreading in different stages. For example, the fire spread probability may include a probability of spreading to 10 square kilometers of 92%, a probability of spreading to 30 square kilometers of 67%, a probability of spreading to 19 square kilometers of 34%, etc. The embodiment of the present invention does not impose too many restrictions on this; the characteristic mutation data is matched with the fire spread probability to obtain sample data, and the above sample data is input into the second fire warning training model to obtain the trained second fire warning training model.

[0080] Specifically, the step of inputting the characteristic mutation data and the corresponding pre-divided fire diffusion probability into the second fire warning training model to obtain the trained second fire warning training model includes:

[0081] Converting the characteristic mutation data and the corresponding pre-divided fire diffusion probability into a diffusion mutation characteristic vector;

[0082] Converting the diffusion mutation feature vector into second mutation dimension data;

[0083] The mutation dimension data is input into the input layer, hidden layer, activation layer, and output layer of the Bayesian neural network to obtain a trained Bayesian neural network.

[0084] Specifically, the second fire warning training model can be a Bayesian neural network. It should be noted that other types of neural networks can also be used to train the corresponding fire warning training model based on the sample data composed of the fire spread probability. The embodiment of the present invention does not impose too many restrictions on this. The structure of the Bayesian neural network may include an input layer, a hidden layer, an activation layer, and an output layer. The embodiment of the present invention does not impose too many restrictions on the structure of the Bayesian neural network. Through the use of the Bayesian neural network, the prediction accuracy of the fire spread probability is improved.

[0085] In another preferred embodiment of the embodiments of the present invention, the method includes: inputting the characteristic mutation data and the corresponding pre-divided fire hazard data into the third fire warning training model to obtain the trained third fire warning training model; inputting the third fire data into the trained third fire warning training model to obtain output specific fire hazard data.

[0086] The fire hazard data may refer to the degree of loss after a fire occurs. For example, the fire hazard data may be divided into extremely large losses, major losses, general losses, etc. It may also be divided into other fire hazard data. The embodiment of the present invention does not impose too many restrictions on this. The above-mentioned fire hazard data and characteristic mutation data constitute sample data and are input into the third fire warning training model to obtain the trained third fire warning training model.

[0087] It should be noted that the third fire warning training model can be multiple types of supervised machine learning models, and the embodiment of the present invention does not impose too many restrictions on this.

[0088] The second fire data and the third fire data may also be data such as images, sounds, temperatures or air water vapor content of an area containing various flammable objects, and the embodiment of the present invention does not impose excessive restrictions on this.

[0089] After obtaining the trained second fire warning training model or the trained third fire warning training model, the second fire data or the third fire data may be input into the model to obtain output specific fire spread probability data or specific fire hazard data.

[0090] In an embodiment of the present invention, fire history data is obtained; wherein the fire history data includes fire history image data, fire history sound data, fire history temperature data, and water vapor evaporation degree data; classification features of the fire history data are extracted, and feature mutation data are screened out from the classification features; the feature mutation data and the corresponding pre-divided fire warning types are input into the first fire warning training model to obtain a trained first fire warning training model; the first fire data is input into the trained first fire warning training model to obtain output specific fire warning type data; the mutation features between different fire types and different fire intensities are trained to obtain feature data between different fire intensities and fire types, thereby improving the accuracy, timeliness and sensitivity of fire identification, and improving the diversity and accuracy of fire warnings.

[0091] It should be noted that, for the method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that this embodiment is not limited by the order of the actions described, because according to this embodiment, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily required by this embodiment.

[0092] Reference Figure 2 , shows a structural block diagram of an embodiment of a fire warning device based on an intelligent model of this embodiment, which may specifically include the following modules:

[0093] The fire history data acquisition module 301 is used to acquire fire history data; wherein the fire history data includes fire history image data, fire history sound data, fire history temperature data, and water vapor evaporation degree data;

[0094] An extraction module 302 is used to extract classification features of the fire history data and filter out feature mutation data from the classification features;

[0095] A first training module 303, used for inputting the characteristic mutation data and the corresponding pre-classified fire warning types into the first fire warning training model to obtain a trained first fire warning training model;

[0096] The first output module 304 is used to input the first fire data into the trained first fire warning training model to obtain output specific fire warning type data.

[0097] Preferably, the device comprises:

[0098] A second training module, used for inputting the characteristic mutation data and the corresponding pre-divided fire diffusion probability into the second fire warning training model to obtain a trained second fire warning training model;

[0099] The second output module is used to input the second fire data into the trained second fire warning training model to obtain output specific fire spread probability data.

[0100] Preferably, the device comprises:

[0101] A third training module, used for inputting the characteristic mutation data and the corresponding pre-divided fire hazard data into the third fire warning training model to obtain a trained third fire warning training model;

[0102] The third output module is used to input the third fire data into the trained third fire warning training model to obtain output specific fire hazard data.

[0103] Preferably, the extraction module comprises:

[0104] A first extraction submodule is used to extract image classification data of the fire history image data;

[0105] A second extraction submodule is used to extract the sound classification data of the fire history sound data;

[0106] A third extraction submodule is used to extract temperature classification data of the fire history temperature data;

[0107] The fourth extraction submodule is used to extract the water vapor evaporation classification data from the water vapor evaporation degree data.

[0108] Preferably, the extraction module comprises:

[0109] A fifth extraction submodule, used to extract image mutation classification features from the image classification data;

[0110] A sixth extraction submodule, used to extract sound mutation classification features from the sound classification data;

[0111] A seventh extraction submodule, used to extract temperature mutation classification features from the temperature classification data;

[0112] The eighth extraction submodule is used to extract water vapor evaporation mutation classification features from the water vapor evaporation classification data.

[0113] Preferably, the first training module includes:

[0114] A first conversion submodule, used to convert the characteristic mutation data and the corresponding pre-divided fire warning types into a warning mutation characteristic vector;

[0115] A second conversion submodule, used to convert the warning mutation feature vector into first mutation dimension data;

[0116] The first training submodule is used to input the mutation dimension data into the convolution layer, pooling layer, fully connected layer, and activation layer of the first fire warning training model to obtain the trained first fire warning training model.

[0117] Preferably, the second fire warning training model includes a Bayesian neural network; and the first training module includes:

[0118] A third conversion submodule is used to convert the characteristic mutation data and the corresponding pre-divided fire diffusion probability into a diffusion mutation characteristic vector;

[0119] A fourth conversion submodule, used for converting the diffusion mutation feature vector into second mutation dimension data;

[0120] The second training submodule is used to input the mutation dimension data into the input layer, hidden layer, activation layer, and output layer of the Bayesian neural network to obtain a trained Bayesian neural network.

[0121] Each module in the fire warning device based on the intelligent model can be implemented in whole or in part by software, hardware and their combination. Each module can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the corresponding operations of each module.

[0122] The fire warning device based on the intelligent model provided above can be used to execute the fire warning method based on the intelligent model provided in any of the above embodiments, and has corresponding functions and beneficial effects.

[0123] In one embodiment, a computer device is provided, whose internal structure diagram can be as follows: Figure 3 As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a fire warning device method based on an intelligent model is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a key, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0124] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0125] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the steps of the above embodiment are implemented:

[0126] Obtain fire history data;

[0127] The fire history data includes fire history image data, fire history sound data, fire history temperature data, and water vapor evaporation degree data;

[0128] Extracting classification features of the fire history data, and filtering out feature mutation data from the classification features;

[0129] Inputting the characteristic mutation data and the corresponding pre-classified fire warning types into the first fire warning training model to obtain a trained first fire warning training model;

[0130] The first fire data is input into the trained first fire warning training model to obtain output specific fire warning type data.

[0131] Preferably, the method comprises:

[0132] Inputting the characteristic mutation data and the corresponding pre-divided fire diffusion probability into the second fire warning training model to obtain a trained second fire warning training model;

[0133] The second fire data is input into the trained second fire warning training model to obtain output specific fire spread probability data.

[0134] Preferably, the method comprises:

[0135] Inputting the characteristic mutation data and the corresponding pre-divided fire hazard data into the third fire warning training model to obtain a trained third fire warning training model;

[0136] The third fire data is input into the trained third fire warning training model to obtain output specific fire hazard data.

[0137] Preferably, extracting classification features of the fire history data and screening out feature mutation data from the classification features includes:

[0138] Extracting image classification data of the fire history image data;

[0139] and / or, extracting sound classification data of the fire history sound data;

[0140] and / or, extracting temperature classification data of the fire history temperature data;

[0141] And / or, extracting water vapor evaporation classification data from the water vapor evaporation degree data.

[0142] Preferably, extracting classification features of the fire history data and screening out feature mutation data from the classification features includes:

[0143] Extracting image mutation classification features from the image classification data;

[0144] and / or, extracting sound mutation classification features from the sound classification data;

[0145] and / or, extracting temperature mutation classification features from the temperature classification data;

[0146] And / or, extracting water vapor evaporation mutation classification features from the water vapor evaporation classification data.

[0147] Preferably, the step of inputting the characteristic mutation data and the corresponding pre-classified fire warning types into the first fire warning training model to obtain the trained first fire warning training model includes:

[0148] Converting the characteristic mutation data and the corresponding pre-divided fire warning types into a warning mutation characteristic vector;

[0149] Converting the early warning mutation feature vector into first mutation dimension data;

[0150] The mutation dimension data is input into the convolution layer, pooling layer, fully connected layer and activation layer of the first fire warning training model to obtain the trained first fire warning training model.

[0151] Preferably, the second fire warning training model includes a Bayesian neural network; the step of inputting the characteristic mutation data and the corresponding pre-divided fire spread probability into the second fire warning training model to obtain the trained second fire warning training model includes:

[0152] Converting the characteristic mutation data and the corresponding pre-divided fire diffusion probability into a diffusion mutation characteristic vector;

[0153] Converting the diffusion mutation feature vector into second mutation dimension data;

[0154] The mutation dimension data is input into the input layer, hidden layer, activation layer, and output layer of the Bayesian neural network to obtain a trained Bayesian neural network.

[0155] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above embodiment are implemented:

[0156] Obtain fire history data;

[0157] The fire history data includes fire history image data, fire history sound data, fire history temperature data, and water vapor evaporation degree data;

[0158] Extracting classification features of the fire history data, and filtering out feature mutation data from the classification features;

[0159] Inputting the characteristic mutation data and the corresponding pre-classified fire warning types into the first fire warning training model to obtain a trained first fire warning training model;

[0160] The first fire data is input into the trained first fire warning training model to obtain output specific fire warning type data.

[0161] Preferably, the method comprises:

[0162] Inputting the characteristic mutation data and the corresponding pre-divided fire diffusion probability into the second fire warning training model to obtain a trained second fire warning training model;

[0163] The second fire data is input into the trained second fire warning training model to obtain output specific fire spread probability data.

[0164] Preferably, the method comprises:

[0165] Inputting the characteristic mutation data and the corresponding pre-divided fire hazard data into the third fire warning training model to obtain a trained third fire warning training model;

[0166] The third fire data is input into the trained third fire warning training model to obtain output specific fire hazard data.

[0167] Preferably, extracting classification features of the fire history data and screening out feature mutation data from the classification features includes:

[0168] Extracting image classification data of the fire history image data;

[0169] and / or, extracting sound classification data of the fire history sound data;

[0170] and / or, extracting temperature classification data of the fire history temperature data;

[0171] And / or, extracting water vapor evaporation classification data from the water vapor evaporation degree data.

[0172] Preferably, extracting classification features of the fire history data and screening out feature mutation data from the classification features includes:

[0173] Extracting image mutation classification features from the image classification data;

[0174] and / or, extracting sound mutation classification features from the sound classification data;

[0175] and / or, extracting temperature mutation classification features from the temperature classification data;

[0176] And / or, extracting water vapor evaporation mutation classification features from the water vapor evaporation classification data.

[0177] Preferably, the step of inputting the characteristic mutation data and the corresponding pre-classified fire warning types into the first fire warning training model to obtain the trained first fire warning training model includes:

[0178] Converting the characteristic mutation data and the corresponding pre-divided fire warning types into a warning mutation characteristic vector;

[0179] Converting the early warning mutation feature vector into first mutation dimension data;

[0180] The mutation dimension data is input into the convolution layer, pooling layer, fully connected layer and activation layer of the first fire warning training model to obtain the trained first fire warning training model.

[0181] Preferably, the second fire warning training model includes a Bayesian neural network; the step of inputting the characteristic mutation data and the corresponding pre-divided fire spread probability into the second fire warning training model to obtain the trained second fire warning training model includes:

[0182] Converting the characteristic mutation data and the corresponding pre-divided fire diffusion probability into a diffusion mutation characteristic vector;

[0183] Converting the diffusion mutation feature vector into second mutation dimension data;

[0184] The mutation dimension data is input into the input layer, hidden layer, activation layer, and output layer of the Bayesian neural network to obtain a trained Bayesian neural network.

[0185] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0186] It will be appreciated by those skilled in the art that the embodiments of the present invention may be provided as methods, devices, or computer program products. Therefore, the embodiments of the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the embodiments of the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0187] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of the methods, terminal devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0188] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0189] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable terminal device. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0190] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present invention.

[0191] Finally, 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 terminal 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 terminal device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or terminal device including the elements.

[0192] The above is a detailed introduction to a fire warning method based on an intelligent model, a fire warning device based on an intelligent model, a computer device and a computer-readable storage medium provided by the present invention. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for general technical personnel in this field, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A fire warning method based on an intelligent model, characterized in that: The method comprises: Obtain fire history data; The fire history data includes fire history image data, fire history sound data, fire history temperature data, and water vapor evaporation degree data; Extracting classification features of the fire history data, and filtering out feature mutation data from the classification features; Inputting the characteristic mutation data and the corresponding pre-classified fire warning types into the first fire warning training model to obtain a trained first fire warning training model; The first fire data is input into the trained first fire warning training model to obtain output specific fire warning type data.

2. The fire early warning method based on intelligent model according to claim 1 is characterized in that: The method comprises: Inputting the characteristic mutation data and the corresponding pre-divided fire diffusion probability into the second fire warning training model to obtain a trained second fire warning training model; The second fire data is input into the trained second fire warning training model to obtain output specific fire spread probability data.

3. The fire early warning method based on intelligent model according to claim 1 is characterized in that: The method comprises: Inputting the characteristic mutation data and the corresponding pre-divided fire hazard data into the third fire warning training model to obtain a trained third fire warning training model; The third fire data is input into the trained third fire warning training model to obtain output specific fire hazard data.

4. The fire early warning method based on intelligent model according to claim 1 is characterized in that: The step of extracting classification features of the fire history data and selecting feature mutation data from the classification features includes: Extracting image classification data of the fire history image data; and / or, extracting sound classification data of the fire history sound data; and / or, extracting temperature classification data of the fire history temperature data; And / or, extracting water vapor evaporation classification data from the water vapor evaporation degree data.

5. The fire early warning method based on intelligent model according to claim 4 is characterized in that: The step of extracting classification features of the fire history data and selecting feature mutation data from the classification features includes: Extracting image mutation classification features from the image classification data; and / or, extracting sound mutation classification features from the sound classification data; and / or, extracting temperature mutation classification features from the temperature classification data; And / or, extracting water vapor evaporation mutation classification features from the water vapor evaporation classification data.

6. The fire early warning method based on intelligent model according to claim 5 is characterized in that: The step of inputting the characteristic mutation data and the corresponding pre-classified fire warning types into the first fire warning training model to obtain the trained first fire warning training model includes: Converting the characteristic mutation data and the corresponding pre-divided fire warning types into a warning mutation characteristic vector; Converting the early warning mutation feature vector into first mutation dimension data; The mutation dimension data is input into the convolution layer, pooling layer, fully connected layer and activation layer of the first fire warning training model to obtain the trained first fire warning training model.

7. The fire early warning method based on intelligent model according to claim 5 is characterized in that: The second fire warning training model includes a Bayesian neural network; the characteristic mutation data and the corresponding pre-divided fire diffusion probability are input into the second fire warning training model to obtain a trained second fire warning training model, including: Converting the characteristic mutation data and the corresponding pre-divided fire diffusion probability into a diffusion mutation characteristic vector; Converting the diffusion mutation feature vector into second mutation dimension data; The mutation dimension data is input into the input layer, hidden layer, activation layer, and output layer of the Bayesian neural network to obtain a trained Bayesian neural network.

8. A fire warning device based on an intelligent model, characterized in that: The device comprises: A fire history data acquisition module, used to acquire fire history data; wherein the fire history data includes fire history image data, fire history sound data, fire history temperature data, and water vapor evaporation degree data; An extraction module, used to extract classification features of the fire history data, and filter out feature mutation data from the classification features; A first training module, used for inputting the characteristic mutation data and the corresponding pre-classified fire warning types into the first fire warning training model to obtain a trained first fire warning training model; The first output module is used to input the first fire data into the trained first fire warning training model to obtain output specific fire warning type data.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the fire warning method based on the intelligent model described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the fire warning method based on an intelligent model as described in any one of claims 1 to 7 are implemented.

Images