An electrical fire monitoring and fire protection Internet of Things system

Abstract

This invention provides an electrical fire monitoring and fire protection IoT system. The system includes a control module and a camera module controlled by the control module. The camera module acquires flame height information and extracts the maximum value, thereby obtaining the movement distance and direction of the maximum value. A two-level attention method is then used to acquire the flame pattern diagram. The beneficial effects of this invention are: it enables the judgment of the fire pattern, allowing for targeted firefighting when using fire-fighting tools, thus preventing greater damage to the area where the equipment is located and greatly protecting the safety of the area.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence, and in particular to an electrical fire monitoring and fire protection Internet of Things system. Background Technology

[0002] In a mechanized society, electrical equipment may operate unsupervised. Under certain conditions, electrical equipment may short-circuit, causing a fire and damaging other electrical appliances.

[0003] Because electrical fires are unpredictable and the source of the fire cannot be located immediately, current technologies may result in insufficient extinguishing agents in the area where the equipment is located. Therefore, it is desirable to limit the amount of extinguishing agent used to extinguish fires by concentrating any release of extinguishing agent on specific areas that require it, rather than spreading it throughout the entire area where the equipment is located. Furthermore, extinguishing agents themselves may be harmful to some types of electrical equipment. Therefore, it is desirable to limit the distribution of extinguishing agents to the location where fire suppression is needed, as the nature of the fire cannot be assessed. However, current technologies only extinguish the fire at the ignition point, which may result in incomplete extinguishing and further damage to the equipment. Summary of the Invention

[0004] The main objective of this invention is to provide an electrical fire monitoring and fire protection Internet of Things system, which aims to solve the problem of being unable to determine the form of a fire.

[0005] This invention provides an electrical fire monitoring and fire protection Internet of Things system, comprising: a control module and a camera module controlled by the control module;

[0006] The camera module includes a first camera in the horizontal shooting direction and a second camera in the vertical shooting direction, used to capture images of the area where the device is located and upload them to the control module;

[0007] When the captured images uploaded to the control module by the first camera and / or the second camera contain firelight, multiple frames of each camera are captured at preset time intervals to obtain a first set of captured frames corresponding to the first camera and a second set of captured frames corresponding to the second camera; wherein, a set of captured frames consists of captured frames from the same camera at different time points.

[0008] Perform wavelet transform on each frame in the first and second set of shooting frames in each group to obtain the corresponding first and second set of photos.

[0009] Obtain the height information of the flames in each photo in the first photo set, and obtain the maximum value of the flame height information and mark the horizontal position information of the maximum value;

[0010] Obtain the flame position information of each photo in the second photo set, and combine it with the horizontal position information to determine the coordinate position of the maximum value in each photo in the second photo set;

[0011] Based on the coordinates of each maximum value, calculate the distance between each maximum value and the maximum value in the nearest previous photo, as well as the burning direction; the burning direction is the direction from the maximum value in the nearest previous photo to the current maximum value.

[0012] A first attention score is set based on each distance to each maximum value, and a second attention score is set based on the direction;

[0013] The weighted sum of the flame height corresponding to each maximum value, as well as the corresponding first attention score and second attention score, is obtained.

[0014] Acquire real-time frames captured by the second camera;

[0015] The real-time frames and weighted sums are input into a preset neural network model to obtain a flame situation map; wherein, the neural network model is trained based on multiple sets of real-time frames and corresponding flame situation maps, and the flame situation map is marked with the danger intensity value of each location.

[0016] Furthermore, it also includes a data acquisition module;

[0017] The data acquisition module obtains a device distribution map of the area where the device is located and uploads it to the control module.

[0018] The control module marks the devices in the real-time frame based on the device distribution map to obtain the marked real-time frame.

[0019] Furthermore, it also includes a temperature sensor array;

[0020] The temperature sensor array is connected to the control module and to the devices in the area where the device is located, for acquiring the temperature of each device;

[0021] When the temperature of the device exceeds the set operating temperature, the data is uploaded to the control module, and the control module controls the camera module to monitor the device.

[0022] Furthermore, it also includes a fire suppression module, which is connected to the control module. The fire suppression module includes at least one fire extinguishing device, and the control module controls the fire suppression module to extinguish the fire according to the flame pattern diagram.

[0023] Furthermore, it also includes an alarm module, which is connected to the control module;

[0024] The control module sends the distance between each maximum value and the maximum value in the nearest previous photo to the alarm module;

[0025] The alarm module is used to sum the various distances to obtain the total displacement.

[0026] Determine whether the total displacement is greater than a preset displacement.

[0027] If the displacement exceeds the preset amount, a distress signal will be sent to the fire department.

[0028] The present invention also provides a fire monitoring method, implemented through the aforementioned fire protection IoT system, comprising:

[0029] When the captured images uploaded to the control module by the first camera and / or the second camera contain firelight, multiple frames of each camera are acquired at preset time intervals to obtain a first set of captured frames corresponding to the first camera and a second set of captured frames corresponding to the second camera; wherein, a set of captured frames consists of captured frames from the same camera at different time points.

[0030] Perform wavelet transform on each frame in the first and second set of shooting frames in each group to obtain the corresponding first and second set of photos.

[0031] Obtain the height information of the flames in each photo in the first photo set, and obtain the maximum value of the flame height information and mark the horizontal position information of the maximum value;

[0032] Obtain the flame position information of each photo in the second photo set, and combine it with the horizontal position information to determine the coordinate position of the maximum value in each photo in the second photo set;

[0033] Based on the coordinates of each maximum value, calculate the distance between each maximum value and the maximum value in the nearest previous photo, as well as the burning direction; the burning direction is the direction from the maximum value in the nearest previous photo to the current maximum value.

[0034] A first attention score is set based on each distance to each maximum value, and a second attention score is set based on the direction;

[0035] The weighted sum of the flame height corresponding to each maximum value, as well as the corresponding first attention score and second attention score, is obtained.

[0036] Acquire real-time frames captured by the second camera;

[0037] The real-time frames and weighted sums are input into a preset neural network model to obtain a flame situation map; wherein, the neural network model is trained based on multiple sets of real-time frames and corresponding flame situation maps, and the flame situation map is marked with the danger intensity value of each location.

[0038] Furthermore, before the step of inputting the real-time frame and the weighted sum into a preset neural network model to obtain the flame pattern diagram, the method further includes:

[0039] Obtain the device distribution map of the area where the device is located and upload it to the control module;

[0040] The control module marks the devices in the real-time frame based on the device distribution map to obtain the marked real-time frame.

[0041] Furthermore, after the step of inputting the real-time frame and the weighted sum into a preset neural network model to obtain a flame pattern diagram, the method further includes:

[0042] A firefighting plan is formulated based on the aforementioned flame situation diagram;

[0043] Firefighting efforts are carried out in the area where the equipment is located based on the firefighting plan and the marked real-time frames.

[0044] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of any of the methods described above.

[0045] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of any of the methods described above.

[0046] The beneficial effects of this invention are as follows: By acquiring the height information of the flame through the camera module and extracting the maximum value, the moving distance and direction of the maximum value can be obtained. Then, by acquiring the flame situation map through a two-level attention method, the fire situation can be judged. When using fire-fighting tools to extinguish the fire, the fire can be extinguished in a targeted manner, thereby avoiding greater damage to the area where the equipment is located and greatly protecting the safety of the area where the equipment is located. Attached Figure Description

[0047] Figure 1 This is a schematic diagram of the structure of an electrical fire monitoring and fire protection Internet of Things system according to an embodiment of the present invention;

[0048] Figure 2 This is a flowchart illustrating a fire monitoring method according to an embodiment of the present invention;

[0049] Figure 3This is a schematic block diagram of the structure of a computer device according to an embodiment of this application.

[0050] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Implementation

[0051] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0052] It should be noted that all directional indicators (such as up, down, left, right, front, back, etc.) in the embodiments of the present invention are only used to explain the relative positional relationship and movement of each component in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indicator will also change accordingly. The connection can be a direct connection or an indirect connection.

[0053] In this article, the term "and / or" is merely a description of the relationship between related objects, indicating that there can be three kinds of relationships. For example, A and B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0054] Furthermore, in this invention, descriptions involving "first," "second," etc., are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first" or "second" may explicitly or implicitly include at least one of that feature. Additionally, the technical solutions of the various embodiments can be combined with each other, but only on the basis of being achievable by those skilled in the art. When the combination of technical solutions is contradictory or impossible to implement, such a combination of technical solutions should be considered non-existent and not within the scope of protection claimed by this invention.

[0055] Reference Figure 1 The present invention proposes an electrical fire monitoring and fire protection Internet of Things system, comprising: a control module 20 and a camera module 10 controlled by the control module 20;

[0056] The control module 20 is a data processing center, which can be hardware with independent computing capabilities, such as a chip or a microcontroller.

[0057] The camera module 10 includes a first camera in the horizontal shooting direction and a second camera in the vertical shooting direction, which are used to capture the area where the device is located and upload the images to the control module 20.

[0058] The locations of the first and second cameras are not limited; they can be used to capture images of the area where the device is located and to pinpoint the location of the flames.

[0059] When the captured images uploaded to the control module by the first camera and / or the second camera contain firelight, multiple frames of each camera are captured at preset time intervals to obtain a first set of captured frames corresponding to the first camera and a second set of captured frames corresponding to the second camera; wherein, a set of captured frames consists of captured frames from the same camera at different time points.

[0060] The method for determining whether there is fire in the images captured by the first and second cameras can be as follows: the captured images are uploaded to the control module 20, and the presence of fire is identified using existing image recognition methods. In some embodiments, a temperature sensor can be used to assist in the determination to enhance accuracy. The preset time interval is not limited and can be set according to specific circumstances; for example, a 2-second interval can be set. It should be noted that each camera captures images, so two photos can be acquired at each interval. The photos (i.e., frames) captured by each individual camera are then combined to obtain the corresponding first and second frame sets.

[0061] Perform wavelet transform on each frame in the first and second set of shooting frames in each group to obtain the corresponding first and second set of photos.

[0062] Since wavelet transform has the characteristics of local decomposition and high compression ratio, it can be used to perform low-frequency processing on the captured frames to obtain the first and second photo sets. Because this application only needs to identify the position of the flame, and other image pixels do not require extensive processing, low-frequency processing can be performed on the captured frames to reduce the amount of data information compared to the original data, in order to avoid data optimization.

[0063] Obtain the height information of the flames in each photo in the first photo set, and obtain the maximum value of the flame height information and mark the horizontal position information of the maximum value;

[0064] The first set of photos can obtain information about the height of the flames. Since the maximum value of the flame can roughly describe the center position of the flame, the shape of the flame can be judged. Therefore, it is necessary to mark the maximum value of the flame height and record the horizontal position information.

[0065] Obtain the flame position information of each photo in the second photo set, and combine it with the horizontal position information to determine the coordinate position of the maximum value in each photo in the second photo set;

[0066] After obtaining the horizontal position information of the maximum flame height, and combining it with the flame position information in each photo in the second photo set, the specific position information of the maximum flame height at each time interval, i.e., the coordinate position, can be obtained.

[0067] Based on the coordinates of each maximum value, calculate the distance between each maximum value and the maximum value in the nearest previous photo, as well as the burning direction; the burning direction is the direction from the maximum value in the nearest previous photo to the current maximum value.

[0068] Since the device generates a combustion direction during the combustion process, which can be determined by the direction of the maximum value, there may be multiple combustion directions. The combustion direction can be determined based on the direction from the maximum value in the nearest previous photo to the current maximum value. The distance calculation process can be determined based on the position of each maximum value.

[0069] A first attention score is set based on each distance to each maximum value, and a second attention score is set based on the direction;

[0070] It should be noted that the maxima mentioned here only refer to the maxima obtained after the second capture. For the first maxima, there is no need to calculate the attention score. The first attention score can be calculated using the formula... Calculate, where, Indicates the first The first attention score corresponding to each maxima Indicates the first The distance between the first maximum and the nearest maximum in the previous image. The second attention score can also be calculated using the softmax function, which will not be elaborated here.

[0071] The weighted sum of the flame height corresponding to each maximum value, as well as the corresponding first attention score and second attention score, is obtained.

[0072] The weighted sum is calculated by summing the second attention score and the corresponding first attention score, and then multiplying it by the first vector to obtain the weighted result for each second target image. Other weighted summation methods are also possible, such as multiplying each attention score by a preset value and then summing them. This can be set according to the actual situation.

[0073] Acquire real-time frames captured by the second camera;

[0074] The real-time frames and weighted sums are input into a preset neural network model to obtain a flame situation map; wherein, the neural network model is trained based on multiple sets of real-time frames and corresponding flame situation maps, and the flame situation map is marked with the danger intensity value of each location.

[0075] After acquiring real-time frames, they are weighted and input into a preset neural network model. Since the real-time frames contain the positions of various devices, the preset neural network model can generate a corresponding flame pattern. The preset neural network model is trained by generating trend vectors from different real-time frames and their corresponding weighted sums. This generation involves converting each real-time frame and its corresponding weighted sum into corresponding numerical values ​​using a preset vector transformation method, and then arranging these numerical values ​​to generate corresponding vectors. Supervised learning is used to train the neural network model. Specifically, the preset neural network model can be one of a convolutional neural network, a recurrent neural network, or a deep neural network. Taking a deep neural network model as an example, it mainly includes an input layer, hidden layers, and an output layer. The first layer is the input layer, the last layer is the output layer, and the intermediate layers are hidden layers. All layers are fully connected; that is, the first... Any neuron in a layer must be connected to the first neuron. Any neuron in a layer is connected. During the training of a deep neural network model, because the output of the deep neural network model is as close as possible to the actual value to be predicted, the current prediction value of the network can be compared with the actual target value. Then, the weight vector of each layer of the neural network model is updated according to the difference between the two (of course, there is usually a process of optimization before the first update, that is, the parameters of each layer in the deep neural network model are pre-configured). For example, if the network's prediction value is too high, the weight vector is adjusted to make it predict lower. This adjustment is continued until the deep neural network model can predict the actual target value or a value that is very close to the actual target value.

[0076] Specifically, the training process can be represented as pre-acquiring different real-time frames and their corresponding weighted sums, as well as a flame situation map. The flame situation map is a manually labeled map based on the flame conditions. The real-time frames are acquired through camera capture, and the weighted sums are calculated based on relevant images. The specific acquisition methods are the same as those used to calculate the weighted sums and real-time frames, and will not be repeated here. The flame situation map contains hazard intensity values ​​at various locations. These hazard intensity values ​​are used to label the flame situation map. After inputting the real-time frames and weighted sums into the input layer, they are converted into corresponding vectors. The hidden layer predicts the hazard intensity value at each location, and the output layer outputs the final result. This final result is then compared with the preset result (i.e., the data in the flame situation map). The parameters in the hidden layer are continuously adjusted using a loss function to achieve the training objective. A higher output value (loss) of the loss function indicates a greater difference; therefore, training the deep neural network model essentially becomes a process of reducing this loss. This allows for the accurate acquisition of a flame situation map, specifically the danger intensity value at each location. The higher the danger intensity value, the more urgent the need for fire suppression. This enables the assessment of the fire's nature and allows for targeted firefighting efforts when using fire-fighting equipment. This helps prevent greater damage to the area where the equipment is located and greatly protects the safety of that area.

[0077] In one embodiment, it further includes a data acquisition module 60;

[0078] The data acquisition module 60 acquires a device distribution map of the area where the device is located and uploads it to the control module 20.

[0079] The control module 20 marks the devices in the real-time frame based on the device distribution map to obtain the marked real-time frame.

[0080] In this embodiment, in order to better facilitate the determination of the danger intensity value at various locations in the area where the equipment is located, the distribution map of the equipment should be obtained and marked in the real-time frame, so as to facilitate the subsequent judgment of the neural network model.

[0081] In one embodiment, a temperature sensor array 30 is also included;

[0082] The temperature sensor array 30 is connected to the control module 20, and the temperature sensor array 30 is connected to the devices in the area where the device is located, for acquiring the temperature of each device;

[0083] When the temperature of the device exceeds the set operating temperature value, the data is uploaded to the control module 20, and the camera module 10 is controlled by the control module 20 to monitor the device.

[0084] In this embodiment, the temperature sensor array 30 is respectively installed at each device to monitor the temperature of each device. When the temperature is greater than the set operating temperature value of the device, it can be determined that there is a fire hazard. This can assist the first camera and the second camera in identifying whether there is fire in the image, thereby achieving the judgment of whether there is fire. In some embodiments, in order to enhance the accuracy of the judgment, a temperature sensor can also be set to assist in the judgment.

[0085] In one embodiment, the fire suppression module 50 is further included. The fire suppression module 50 is connected to the control module 20. The fire suppression module 50 includes at least one fire extinguishing device. The control module 20 controls the fire suppression module 50 to extinguish the fire according to the flame pattern diagram.

[0086] After acquiring the flame situation map, the fire suppression module 50 can perform targeted fire suppression, thereby ensuring the safety of other equipment in the area where the equipment is located. In a preferred embodiment, if the danger intensity value of the location is greater than the set value, it can be determined that the fire cannot be extinguished. Fire suppression can then be carried out based on the danger intensity values ​​of other locations, thereby saving the dosage of fire extinguishing equipment and achieving a better fire suppression effect.

[0087] In one embodiment, an alarm module 40 is also included, which is connected to the control module 20;

[0088] The control module 20 sends the distance between each maximum value and the maximum value in the nearest previous photo to the alarm module 40;

[0089] The alarm module 40 is used to sum the distances to obtain the total displacement.

[0090] Determine whether the total displacement is greater than a preset displacement.

[0091] If the displacement exceeds the preset amount, a distress signal will be sent to the fire department.

[0092] In this embodiment, since the spread of the fire makes the situation on site uncontrollable, the maximum total displacement can be used to judge the spread. The spread can be determined based on the total displacement. When the total displacement is greater than the preset displacement, it can be considered uncontrollable, and a distress signal can be sent to the fire department. The distress signal can include the location information of the area where the equipment is located and the flame situation diagram, so as to facilitate firefighters to carry out firefighting. If the total displacement is less than or equal to the preset displacement, the fire is still controllable, and it can be uploaded to the relevant person in charge for timely firefighting or alarm. This application does not limit this, and it can be determined by the relevant personnel.

[0093] Reference Figure 2The present invention also provides a fire monitoring method, implemented through the aforementioned fire protection IoT system, comprising:

[0094] S1: When the shooting images uploaded to the control module by the first camera and / or the second camera contain firelight, multiple shooting frames of each camera are acquired at preset time intervals to obtain the first shooting frame set corresponding to the first camera and the second shooting frame set corresponding to the second camera; wherein, a set of shooting frames consists of shooting frames of the same camera at different time points.

[0095] S2: Perform wavelet transform on each frame in the first and second set of shooting frames in each group to obtain the corresponding first and second set of photos;

[0096] S3: Obtain the height information of the flames in each photo in the first photo set, and obtain the maximum value of the flame height information and mark the horizontal position information of the maximum value;

[0097] S4: Obtain the flame position information of each photo in the second photo set, and combine it with the horizontal position information to determine the coordinate position of the maximum value in each photo in the second photo set;

[0098] S5: Based on the coordinates of each maximum value, calculate the distance between each maximum value and the maximum value in the nearest previous photo, as well as the burning direction; the burning direction is the direction from the maximum value in the nearest previous photo to the current maximum value;

[0099] S6: Set the first attention score for each maximum value based on each distance, and set the second attention score based on the direction;

[0100] S7: The flame height corresponding to each maximum value, as well as the corresponding first attention score and second attention score, are weighted and summed to obtain the weighted sum of each maximum value;

[0101] S8: Obtain real-time frames captured by the second camera;

[0102] S9: Input the real-time frames and weighted sums into a preset neural network model to obtain a flame situation map; wherein, the neural network model is trained based on multiple sets of real-time frames and corresponding flame situation maps, and the flame situation map is marked with the danger intensity value of each location.

[0103] As described in step S1 above, when the captured images uploaded to the control module by the first camera and / or the second camera contain flames, multiple frames from each camera are acquired at preset time intervals to obtain a first set of captured frames corresponding to the first camera and a second set of captured frames corresponding to the second camera. A set of captured frames consists of captured frames from the same camera at different times. The camera module 10 includes a first camera in a horizontal shooting direction and a second camera in a vertical shooting direction, used to capture images of the area where the device is located and upload them to the control module 20. The positions of the first and second cameras are not limited; they can capture images of the area where the device is located and locate the position of the flames. The method for determining whether the captured images from the first and second cameras contain flames can be to upload the captured images to the control module 20 and identify whether there are flames in the images using existing image recognition methods. In some embodiments, to enhance the accuracy of the judgment, a temperature sensor can also be used to assist in the judgment. Furthermore, the preset time interval is not limited and can be set according to specific circumstances. For example, a 2-second interval can be set. It should be noted that each camera takes pictures, so two pictures can be obtained at each interval. Then, the pictures taken by each individual camera (i.e., the captured frames) are combined to obtain the corresponding first and second captured frame sets. This allows for an accurate fire situation map, i.e., the danger intensity value of each location. The higher the danger intensity value, the more urgent the fire needs to be extinguished. This enables the assessment of the fire situation, allowing for targeted firefighting when using fire-fighting equipment, thereby preventing greater damage to the area where the equipment is located and greatly protecting the safety of the area.

[0104] As described in step S2 above, wavelet transform is performed on each frame in each of the first and second sets of captured frames to obtain the corresponding first and second photo sets. Since wavelet transform has characteristics such as local decomposition and high compression ratio, it can be used to perform low-frequency processing on the captured frames to obtain the first and second photo sets. Because this application only needs to identify the position of the flame, and other image pixels do not require extensive processing, low-frequency processing can be performed on the captured frames to reduce the amount of data information compared to the original data in order to avoid data optimization.

[0105] As described in step S3 above, the height information of the flames in each photo in the first photo set is obtained, and the maximum value of the flame height information is obtained and the horizontal position information of the maximum value is marked. The flame height information can be obtained in the first photo set. Since the maximum value of the flame can roughly describe the center position of the flame, the form of the flame can be judged. Therefore, it is necessary to mark the maximum value of the flame height and record the horizontal position information.

[0106] As described in step S4 above, the flame position information of each photo in the second photo set is obtained, and the coordinate position of the maximum value in each photo in the second photo set is determined by combining the horizontal position information. After obtaining the horizontal position information of the maximum flame height, the specific position information of the maximum flame height at each time interval, i.e., the coordinate position, can be obtained by combining the flame position information in each photo in the second photo set.

[0107] As described in step S5 above, based on the coordinate positions of each maximum value, the distance between each maximum value and the maximum value in the nearest previous photo and the burning direction are calculated; the burning direction is the direction from the maximum value in the nearest previous photo to the current maximum value; since the device will generate a burning direction during the burning process, and this can be determined by the direction of the maximum value, there may be multiple burning directions, and the burning direction can be determined according to the direction from the maximum value in the nearest previous photo to the current maximum value, and the distance calculation process can be determined according to the position of each maximum value.

[0108] As described in step S6 above, a first attention score is set for each maximum value based on each distance, and a second attention score is set based on the direction. It should be noted that each maximum value here only applies to the maximum values ​​obtained after the second shot. For the maximum value obtained the first time, there is no need to calculate the attention score. The first attention score can be calculated using a formula. Calculate, where, Indicates the first The first attention score corresponding to each maxima Indicates the first The distance between the first maximum and the nearest maximum in the previous image. The second attention score can also be calculated using the softmax function, which will not be elaborated here.

[0109] As described in step S7 above, the flame heights corresponding to each maxima, along with the corresponding first and second attention scores, are weighted and summed to obtain a weighted sum of each maxima. The weighted sum is calculated by weighting the second attention score and the corresponding first attention score, and then multiplying it by the first vector to obtain the weighted result of each second target image. Other weighted summing methods can also be used, such as multiplying each attention score by a preset value and then summing them, which can be set according to the actual situation.

[0110] As described in steps S8-S9 above, real-time frames captured by the second camera are acquired; the real-time frames and their weighted sums are input into a preset neural network model to obtain a flame pattern diagram; wherein, the neural network model is trained based on multiple sets of real-time frames and corresponding flame pattern diagrams, and the flame pattern diagrams are marked with the danger intensity values ​​of each location. After acquiring the real-time frames, they are input into the preset neural network model along with their weighted sums. Since the real-time frames contain the location information of each device, the input into the preset neural network model can obtain the corresponding flame pattern diagram. The preset neural network model is trained by generating corresponding trend vectors from different real-time frames and their corresponding weighted sums. The generation method involves converting each real-time frame and its corresponding weighted sum into corresponding result values ​​according to a preset vector transformation method, and arranging the result values ​​to generate corresponding vectors. The neural network model is trained using supervised learning. Specifically, the preset neural network model can be one of a convolutional neural network model, a recurrent neural network model, or a deep neural network model. In the case of a deep neural network model, it mainly includes an input layer, a hidden layer, and an output layer. The first layer is the input layer, the last layer is the output layer, and the layers in between are hidden layers. All layers are fully connected; that is, the first... Any neuron in a layer must be connected to the first neuron. Any neuron in a layer is connected. During the training of a deep neural network model, because the output of the deep neural network model is as close as possible to the actual value to be predicted, the current prediction value of the network can be compared with the actual target value. Then, the weight vector of each layer of the neural network model is updated according to the difference between the two (of course, there is usually a process of optimization before the first update, that is, the parameters of each layer in the deep neural network model are pre-configured). For example, if the network's prediction value is too high, the weight vector is adjusted to make it predict lower. This adjustment is continued until the deep neural network model can predict the actual target value or a value that is very close to the actual target value. Specifically, the training process can be represented as pre-acquiring different real-time frames and their corresponding weighted sums, as well as a flame situation map. The flame situation map is a manually labeled map based on the flame conditions. The real-time frames are acquired through camera capture, and the weighted sums are calculated based on relevant images. The specific acquisition methods are the same as those used to calculate the weighted sums and real-time frames, and will not be repeated here. The flame situation map contains hazard intensity values ​​at various locations. These hazard intensity values ​​are used to label the flame situation map. After inputting the real-time frames and weighted sums into the input layer, they are converted into corresponding vectors. The hidden layer predicts the hazard intensity value at each location, and the output layer outputs the final result. This final result is then compared with the preset result (i.e., the data in the flame situation map). The parameters in the hidden layer are continuously adjusted using a loss function to achieve the training objective. A higher output value (loss) of the loss function indicates a greater difference; therefore, training the deep neural network model essentially becomes a process of reducing this loss. This allows for the accurate acquisition of a flame situation map, specifically the danger intensity value at each location. A higher danger intensity value indicates that the location most urgently needs fire suppression, thus enabling the assessment of the fire's form. When using firefighting equipment, targeted firefighting efforts can be implemented, preventing greater damage to the equipment area and significantly protecting its safety. In one embodiment, before step S9, which involves inputting the real-time frame and its weighted sum into a preset neural network model to obtain the flame situation map, the method further includes:

[0111] S801: Obtain the device distribution map of the area where the device is located and upload it to the control module 20;

[0112] S802: The control module 20 marks the devices in the real-time frame based on the device distribution map to obtain the marked real-time frame.

[0113] In this embodiment, in order to better facilitate the determination of the danger intensity value at various locations in the area where the equipment is located, the distribution map of the equipment should be obtained and marked in the real-time frame, so as to facilitate the subsequent judgment of the neural network model.

[0114] In one embodiment, after step S9 of inputting the real-time frame and the weighted sum into a preset neural network model to obtain a flame pattern diagram, the method further includes:

[0115] S1001: Set a firefighting plan based on the flame situation diagram;

[0116] S1002: Extinguish the fire in the area where the equipment is located according to the firefighting plan and the marked real-time frames.

[0117] After acquiring the flame situation map, the fire suppression module 50 can perform targeted fire suppression, thereby ensuring the safety of other equipment in the area where the equipment is located. In a preferred embodiment, if the danger intensity value of the location is greater than the set value, it can be determined that the fire cannot be extinguished. Fire suppression can then be carried out based on the danger intensity values ​​of other locations, thereby saving the dosage of fire extinguishing equipment and achieving a better fire suppression effect.

[0118] The beneficial effects of this invention are as follows: By acquiring the height information of the flame through the camera module 10 and extracting the maximum value, the moving distance and direction of the maximum value can be obtained. Then, by acquiring the flame situation map through a two-level attention method, the fire situation can be judged. When using fire-fighting tools to extinguish the fire, the fire can be extinguished in a targeted manner, thereby avoiding greater damage to the area where the equipment is located and greatly protecting the safety of the area where the equipment is located.

[0119] Reference Figure 3 This application also provides a computer device, which may be a server, and its internal structure may be as follows: Figure 3 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores various coordinate positions, etc. The network interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it can implement the electrical fire monitoring and fire protection IoT system described in any of the above embodiments.

[0120] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer equipment on which the present application is applied.

[0121] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, can implement the electrical fire monitoring and fire protection IoT system described in any of the above embodiments.

[0122] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media provided in this application and used in the embodiments can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0123] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0124] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0125] Foundational technologies in artificial intelligence generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies mainly encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0126] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of the claims of the present invention.

Claims

1. An electrical fire monitoring and fire protection Internet of Things (IoT) system, characterized in that, include: A control module and a camera module controlled by the control module; The camera module includes a first camera in the horizontal shooting direction and a second camera in the vertical shooting direction, used to capture images of the area where the device is located and upload them to the control module; When the captured images uploaded to the control module by the first camera and / or the second camera contain firelight, multiple frames of each camera are acquired at preset time intervals to obtain a first set of captured frames corresponding to the first camera and a second set of captured frames corresponding to the second camera; wherein, a set of captured frames consists of captured frames from the same camera at different time points. Perform wavelet transform on each frame in the first and second set of shooting frames in each group to obtain the corresponding first and second set of photos. Obtain flame height information from each photo in the first photo set, and obtain the maximum value of the flame height information and mark the horizontal position information of the maximum value; Obtain the flame position information of each photo in the second photo set, and combine it with the horizontal position information to determine the coordinate position of the maximum value in each photo in the second photo set; Based on the coordinates of each maximum value, calculate the distance between each maximum value and the maximum value in the nearest previous photo, as well as the burning direction; the burning direction is the direction from the maximum value in the nearest previous photo to the current maximum value. A first attention score is set based on each distance to each maximum value, and a second attention score is set based on the direction; The weighted sum of the flame height corresponding to each maximum value, as well as the corresponding first attention score and second attention score, is obtained. Acquire real-time frames captured by the second camera; The real-time frames and weighted sums are input into a preset neural network model to obtain a flame situation map; wherein, the neural network model is trained based on multiple sets of real-time frames and corresponding flame situation maps, and the flame situation map is marked with the danger intensity value of each location.

2. The electrical fire monitoring and fire protection IoT system as described in claim 1, characterized in that, It also includes a data acquisition module; The data acquisition module obtains a device distribution map of the area where the device is located and uploads it to the control module. The control module marks the devices in the real-time frame based on the device distribution map to obtain the marked real-time frame.

3. The electrical fire monitoring and fire protection IoT system as described in claim 1, characterized in that, It also includes a temperature sensor array; The temperature sensor array is connected to the control module and to the devices in the area where the device is located, for acquiring the temperature of each device; When the temperature of the device exceeds the set operating temperature, the data is uploaded to the control module, and the control module controls the camera module to monitor the device.

4. The electrical fire monitoring and fire protection IoT system as described in claim 1, characterized in that, It also includes a fire suppression module, which is connected to the control module. The fire suppression module includes at least one fire extinguishing device, and the control module controls the fire suppression module to extinguish the fire according to the flame pattern diagram.

5. The electrical fire monitoring and fire protection IoT system as described in claim 1, characterized in that, It also includes an alarm module, which is connected to the control module; The control module sends the distance between each maximum value and the maximum value in the nearest previous photo to the alarm module; The alarm module is used to sum the various distances to obtain the total displacement. Determine whether the total displacement is greater than a preset displacement. If the displacement exceeds the preset amount, a distress signal will be sent to the fire department.

6. A fire monitoring method, implemented through the fire protection Internet of Things system according to any one of claims 1-5, characterized in that, include: When the captured images uploaded to the control module by the first camera and / or the second camera contain firelight, multiple frames of each camera are acquired at preset time intervals to obtain a first set of captured frames corresponding to the first camera and a second set of captured frames corresponding to the second camera; wherein, a set of captured frames consists of captured frames from the same camera at different time points. Perform wavelet transform on each frame in the first and second set of shooting frames in each group to obtain the corresponding first and second set of photos. Obtain the height information of the flames in each photo in the first photo set, and obtain the maximum value of the flame height information and mark the horizontal position information of the maximum value; Obtain the flame position information of each photo in the second photo set, and combine it with the horizontal position information to determine the coordinate position of the maximum value in each photo in the second photo set; Based on the coordinates of each maximum value, calculate the distance between each maximum value and the maximum value in the nearest previous photo, as well as the burning direction; the burning direction is the direction from the maximum value in the nearest previous photo to the current maximum value. A first attention score is set based on each distance to each maximum value, and a second attention score is set based on the direction; The weighted sum of the flame height corresponding to each maximum value, as well as the corresponding first attention score and second attention score, is obtained. Acquire real-time frames captured by the second camera; The real-time frames and weighted sums are input into a preset neural network model to obtain a flame situation map; wherein, the neural network model is trained based on multiple sets of real-time frames and corresponding flame situation maps, and the flame situation map is marked with the danger intensity value of each location.

7. The fire monitoring method as described in claim 6, characterized in that, Before the step of inputting the real-time frame and the weighted sum into a preset neural network model to obtain a flame pattern diagram, the method further includes: Obtain the device distribution map of the area where the device is located and upload it to the control module; The control module marks the devices in the real-time frame based on the device distribution map to obtain the marked real-time frame.

8. The fire monitoring method as described in claim 7, characterized in that, After the step of inputting the real-time frame and the weighted sum into a preset neural network model to obtain a flame pattern diagram, the method further includes: A firefighting plan is formulated based on the aforementioned flame situation diagram; Firefighting efforts are carried out in the area where the equipment is located based on the firefighting plan and the marked real-time frames.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 6 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 6 to 8.

Images