Fire extinguishing control method and system based on fire extinguishing patch

Abstract

The application discloses a fire extinguishing control method and system based on a fire extinguishing sticker. The method comprises the following steps: determining whether a potential fire source appears in a target area according to initial image data of the target area at a current time; if the potential fire source appears, determining whether a fire point exists in the target area according to temperature distribution data of the target area; when the fire point exists, determining a working fire extinguishing sticker and a first spraying angle of the working fire extinguishing sticker according to the fire point, wherein the working fire extinguishing sticker is at least one of a plurality of fire extinguishing stickers arranged in the target area; driving the working fire extinguishing sticker to spray fire extinguishing particles at the fire point at the first spraying angle; obtaining a motion trajectory of the fire extinguishing particles sprayed at the first spraying angle, and determining a second spraying angle according to the motion trajectory, the position of the working fire extinguishing sticker and the position of the fire point; and driving the working fire extinguishing sticker to spray fire extinguishing particles at the fire point at the second spraying angle. The application can accurately control the fire extinguishing sticker to perform fire extinguishing treatment.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a fire extinguishing control method and system based on fire extinguishing patches. Background Technology

[0002] In archives and similar institutions that store large amounts of paper documents, fire prevention and firefighting safety measures are of paramount importance to prevent the destruction of valuable materials and the resulting significant economic losses. Since paper documents are damaged by water and difficult to repair, water spraying has been ruled out as a method of fire suppression. Instead, fire extinguishing patches are used to spray extinguishing particles onto the fire to completely extinguish the combustion reaction.

[0003] However, current fire extinguishing patches lack sufficient precision and cannot accurately control the application of the patches to extinguish the burning material at the ignition point, making it difficult to effectively protect important documents in archives during a fire. Summary of the Invention

[0004] To address the aforementioned technical problems, this application proposes a fire extinguishing control method and system based on fire extinguishing patches.

[0005] This application provides a fire extinguishing control method based on fire extinguishing patches, including:

[0006] Based on the initial image data of the target area at the current moment, determine whether there is a potential fire source in the target area;

[0007] If a potential fire source is found, the temperature distribution data of the target area is obtained to determine whether there is an ignition point in the target area. When there is an ignition point, the working fire extinguishing patch and its first spray angle are determined according to the ignition point. The working fire extinguishing patch is at least one of several fire extinguishing patches set in the target area.

[0008] The working fire extinguishing patch sprays fire extinguishing particles toward the fire at a first spray angle;

[0009] The trajectory of the extinguishing particles ejected at the first spray angle is obtained, and the second spray angle is determined based on the trajectory, the position of the working extinguishing patch and the position of the ignition point. The working extinguishing patch is then driven to spray the extinguishing particles toward the ignition point at the second spray angle.

[0010] In one embodiment, the method further includes: determining whether the fire at the ignition point has been extinguished based on the obtained temperature change data and smoke concentration data at the ignition point; if the fire at the ignition point has not been extinguished, and the current temperature value of the ignition point indicated by the temperature change data is greater than a preset temperature, then driving at least one fire extinguishing patch adjacent to the working fire extinguishing patch among a plurality of fire extinguishing patches to spray fire extinguishing particles toward the ignition point, or driving the output end of the inert gas generator corresponding to the ignition point to release inert gas toward the ignition point.

[0011] In one embodiment, after driving at least one fire extinguishing patch adjacent to the working fire extinguishing patch in a plurality of fire extinguishing patches to spray fire extinguishing particles toward the ignition point, or driving the output end of the inert gas generator corresponding to the ignition point to release inert gas toward the ignition point, the method further includes: acquiring the real-time temperature value and real-time smoke concentration of the ignition point to determine whether the fire at the ignition point has been extinguished; if the fire at the ignition point has been extinguished, driving the ventilation system corresponding to the target area to work.

[0012] In one embodiment, determining whether a potential fire source exists in the target area based on the acquired initial image data of the target area at the current moment includes: acquiring the initial image data of the target area at the current moment; performing image preprocessing on the initial image data using an image processing algorithm to obtain image data to be tested; extracting image features from the image data to be tested to obtain feature vectors of the image data to be tested, and inputting the feature vectors into a pre-trained feature classification model to obtain multiple types of image features; if the multiple types of image features include fire light wave features, then a potential fire source is determined to exist; wherein, the fire light wave features are determined based on historical fire image data.

[0013] In one embodiment, the parameters of the image processing algorithm are adjusted through the following steps: acquiring historical image data of historical fire events in the target area and historical illumination data corresponding to the historical fire events; analyzing the historical image data and historical illumination data to determine the correlation between the image judgment deviation of historical fire events and the illumination change pattern of the target area, so as to train the artificial intelligence model and obtain the correlation model; inputting the illumination information of the target area at the current moment into the correlation model to obtain the image judgment deviation corresponding to the illumination information at the current moment, and adjusting the parameters of the image processing algorithm according to the image judgment deviation of the illumination information at the current moment.

[0014] In one embodiment, determining whether there is an ignition point in the target area based on the obtained temperature distribution data of the target area includes: obtaining the temperature distribution data of the target area and performing data preprocessing on the temperature distribution data to obtain a temperature distribution map of the target area; calculating the temperature gradient value at each location in the temperature distribution map and determining the location where the temperature gradient value is greater than a preset gradient threshold as the ignition location; if the area of ​​the region formed by multiple consecutive ignition locations is greater than a preset area threshold, then it is determined that there is an ignition point in the target area.

[0015] In one embodiment, the step of determining the working fire extinguishing patch and its first spray angle based on the ignition point includes: determining the location of the ignition point; determining at least one fire extinguishing patch from a plurality of fire extinguishing patches that is closest to the location of the ignition point as the working fire extinguishing patch; determining the location of the working fire extinguishing patch, and determining the first spray angle based on the location of the working fire extinguishing patch and the location of the ignition point.

[0016] In one embodiment, the step of acquiring the motion trajectory of fire extinguishing particles sprayed at a first spray angle and determining a second spray angle based on the motion trajectory, the position of the working fire extinguishing patch, and the position of the ignition point includes: acquiring a motion sequence image of the fire extinguishing particles sprayed at the first spray angle using a high-speed camera; determining the motion trajectory based on the motion sequence image and obtaining the position of each frame corresponding to the motion trajectory in the motion sequence image; determining the expected motion trajectory of the fire extinguishing particles based on the position of the working fire extinguishing patch and the position of the ignition point; for each frame position, if the difference between the frame position and the corresponding point in the expected motion trajectory is greater than a preset deviation threshold, then determining the difference as the target difference value corresponding to the frame position; inputting all target differences into a pre-trained parameter adjustment model to obtain the parameter adjustment value output by the parameter adjustment model, and determining the second spray angle based on the parameter adjustment value and the first spray angle.

[0017] In one embodiment, determining whether a fire at an ignition point has been extinguished based on the acquired temperature change data and smoke concentration data at the ignition point includes: acquiring temperature change data and smoke concentration data at the ignition point; and within a preset time period after the working fire extinguishing patch has finished spraying, if the current temperature value indicated by the temperature change data is less than a preset temperature threshold and the current smoke concentration value indicated by the smoke concentration data is less than a preset concentration threshold, then determining that the fire at the ignition point has been extinguished.

[0018] This application provides a fire extinguishing control system based on fire extinguishing stickers, the system comprising:

[0019] The potential ignition source detection module is used to determine whether there is a potential ignition source in the target area based on the initial image data of the target area at the current moment.

[0020] The fire extinguishing patch determination module is used to determine whether there is an ignition point in the target area based on the temperature distribution data of the target area if a potential fire source is found. When there is an ignition point, the working fire extinguishing patch and its first spray angle are determined based on the ignition point. The working fire extinguishing patch is at least one of several fire extinguishing patches set in the target area.

[0021] The fire extinguishing patch driving module is used to drive the working fire extinguishing patch to spray fire extinguishing particles toward the fire point at a first spray angle.

[0022] The spray angle adjustment module is used to acquire the movement trajectory of the fire extinguishing particles sprayed at the first spray angle, and determine the second spray angle based on the movement trajectory, the position of the working fire extinguishing patch and the position of the ignition point, and drive the working fire extinguishing patch to spray fire extinguishing particles toward the ignition point at the second spray angle.

[0023] The aforementioned fire extinguishing control method and system based on fire extinguishing patches first determines whether a potential fire source exists in the target area based on the initial image data of the target area at the current moment. If a potential fire source exists, the system determines whether an ignition point exists in the target area based on the temperature distribution data of the target area. If an ignition point exists, the system determines the working fire extinguishing patch and its first spray angle based on the ignition point, wherein the working fire extinguishing patch is at least one of several fire extinguishing patches placed in the target area. Then, the system drives the working fire extinguishing patch to spray fire extinguishing particles toward the ignition point at the first spray angle. The system acquires the trajectory of the fire extinguishing particles sprayed at the first spray angle, and determines the second spray angle based on the trajectory, the position of the working fire extinguishing patch, and the position of the ignition point. The system then drives the working fire extinguishing patch to spray fire extinguishing particles toward the ignition point at the second spray angle. This application uses data acquired by multiple sensors to accurately determine the occurrence of potential fires in the target area. Then, it determines the working fire extinguishing patch and its spray angle based on the ignition point. By analyzing the movement trajectory of the fire extinguishing particles, it adjusts the spray angle of the working fire extinguishing patch to control the fire extinguishing patch to accurately spray the inactivating particles towards the ignition point. This allows for precise control of the fire extinguishing patch to extinguish the burning material at the ignition point, effectively protecting important documents in the archives during a fire. Attached Figure Description

[0024] Figure 1 This is a flowchart illustrating a fire extinguishing control method based on fire extinguishing patches in one embodiment;

[0025] Figure 2 This is a schematic diagram of a fire extinguishing control system based on fire extinguishing stickers in one embodiment. Detailed Implementation

[0026] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0027] In the description of this application, the terms "first," "second," "third," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined with "first," "second," "third," etc., may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more. In the description of this application, the term "comprising" and its variations are open-ended, meaning "including but not limited to." The term "based on" means "at least partially based on." The term "according to" means "at least partially according to." The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments."

[0028] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0029] In the description of this application, it should be noted that, unless otherwise defined, all technical and scientific terms used in this application have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this application is for the purpose of describing specific embodiments only and is not intended to limit the application. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0030] See Figure 1 This document illustrates a flowchart of a fire extinguishing control method based on a fire extinguishing patch, according to an embodiment of this application. The method includes steps S101-S104, and includes:

[0031] Step S101: Based on the initial image data of the target area at the current moment, determine whether there is a potential fire source in the target area.

[0032] Step S102: If a potential fire source is found, the temperature distribution data of the target area is obtained to determine whether there is an ignition point in the target area. When there is an ignition point, the working fire extinguishing patch and its first spray angle are determined according to the ignition point. The working fire extinguishing patch is at least one of several fire extinguishing patches set in the target area.

[0033] Specifically, the target area is continuously monitored by acquiring images using cameras or sensors. By analyzing the initial image data of the target area at the current moment, it is determined whether a potential ignition source exists within the target area. If a potential ignition source is found, the temperature sensors within the target area are activated to collect temperature distribution data. Then, based on the acquired temperature distribution data, it is determined whether an ignition point exists within the target area. If an ignition point exists, at least one of several fire extinguishing pads placed in the target area is designated as a working fire extinguishing pad based on the location coordinates of the ignition point. Finally, the first spray angle of the working fire extinguishing pad is determined based on the location coordinates of the ignition point.

[0034] It should be noted that the location coordinates of the ignition point in the temperature distribution data are two-dimensional coordinates. These need to be calculated using a three-dimensional spatial model of the target area to determine the three-dimensional spatial coordinates of the ignition point. The fire extinguishing patch location information database contains the ID numbers of several fire extinguishing patches placed in the target area and their corresponding three-dimensional spatial coordinates. The spray angle includes both pitch and azimuth angles.

[0035] Step S103: Drive the working fire extinguishing patch to spray fire extinguishing particles toward the fire point at the first spray angle.

[0036] Step S104: Obtain the trajectory of the extinguishing particles sprayed at the first spray angle, and determine the second spray angle based on the trajectory, the position of the working extinguishing patch and the position of the ignition point, and drive the working extinguishing patch to spray the extinguishing particles toward the ignition point at the second spray angle.

[0037] Specifically, the motor drive system of the working fire extinguishing patch's extinguishing particle launching device adjusts the direction of the launching tube according to the first spray angle, aligning the launching tube with the fire point. After the angle adjustment is complete, the working fire extinguishing patch's built-in control unit triggers the extinguishing particle launching device, spraying the extinguishing particles along the direction of the launching tube. It should be noted that the first spray angle can also be calculated by the working fire extinguishing patch's built-in control unit based on the fire point's location coordinates. In this case, a command containing the fire point's location information needs to be sent to the fire extinguishing patch.

[0038] During their flight, the extinguishing particles are subject to gravity, forming a parabolic trajectory before finally landing near the fire point and releasing an ignition agent to extinguish the fire. A high-speed camera is placed along the path of the extinguishing particles to capture their trajectory when sprayed at a first spray angle. Based on the trajectory, the position of the working extinguishing patch, and the location of the fire point, the deviation between the particle landing point and the fire point is analyzed to determine a second spray angle. The working extinguishing patch is then driven to spray the extinguishing particles towards the fire point at this second spray angle.

[0039] The aforementioned fire extinguishing control method based on fire extinguishing patches first determines whether a potential fire source exists in the target area based on the initial image data of the target area at the current moment. If a potential fire source exists, it determines whether an ignition point exists in the target area based on the temperature distribution data of the target area. When an ignition point exists, it determines the working fire extinguishing patch and its first spray angle based on the ignition point, wherein the working fire extinguishing patch is at least one of several fire extinguishing patches placed in the target area. Then, it drives the working fire extinguishing patch to spray fire extinguishing particles toward the ignition point at the first spray angle. It acquires the trajectory of the fire extinguishing particles sprayed at the first spray angle and determines the second spray angle based on the trajectory, the position of the working fire extinguishing patch, and the position of the ignition point, and drives the working fire extinguishing patch to spray fire extinguishing particles toward the ignition point at the second spray angle. This application uses data acquired by multiple sensors to accurately determine the occurrence of potential fires in the target area. Then, it determines the working fire extinguishing patch and its spray angle based on the ignition point. By analyzing the movement trajectory of the fire extinguishing particles, it adjusts the spray angle of the working fire extinguishing patch to control the fire extinguishing patch to accurately spray the inactivating particles towards the ignition point. This allows for precise control of the fire extinguishing patch to extinguish the burning material at the ignition point, effectively protecting important documents in the archives during a fire.

[0040] In one embodiment, the method further includes:

[0041] Based on the obtained temperature change data and smoke concentration data at the ignition point, it is determined whether the fire at the ignition point has been extinguished. If the fire at the ignition point has not been extinguished, and the current temperature value of the ignition point indicated by the temperature change data is greater than the preset temperature, then at least one fire extinguishing patch adjacent to the working fire extinguishing patch among several fire extinguishing patches is driven to spray fire extinguishing particles toward the ignition point, or the output end of the inert gas generator corresponding to the ignition point is driven to release inert gas toward the ignition point.

[0042] Specifically, after the extinguishing particles are sprayed, a thermal imager acquires temperature change data at the ignition point, and a smoke sensor collects smoke concentration data. Based on the acquired temperature change data and smoke concentration data, it is determined whether the fire at the ignition point has been extinguished. If the fire at the ignition point has not been extinguished, and the current temperature value indicated by the temperature change data is greater than the preset temperature, it is determined that there is still a risk of reignition at the ignition point, and a secondary fire suppression procedure must be automatically activated. The secondary fire suppression procedure includes driving at least one extinguishing patch adjacent to the working extinguishing patch to spray extinguishing particles towards the ignition point, or driving the output end of an inert gas generator corresponding to the ignition point to release inert gas towards the ignition point. After the inert gas generator is activated, it first cuts off the air supply to the protected area and releases nitrogen or carbon dioxide inert gas within the area to prevent combustibles from continuing to burn.

[0043] In this embodiment, by analyzing the temperature change data and smoke concentration data at the ignition point, it is determined whether the fire at the ignition point has been extinguished. If the fire has not been extinguished and there is a risk of reignition, a secondary fire extinguishing procedure is taken to ensure that the fire is completely resolved, thereby further improving the effectiveness of the fire extinguishing and control method of this application.

[0044] In one embodiment, after the steps of driving at least one fire extinguishing patch adjacent to the working fire extinguishing patch in a plurality of fire extinguishing patches to spray fire extinguishing particles toward the ignition point, or driving the output end of an inert gas generator corresponding to the ignition point to release inert gas toward the ignition point, the method further includes:

[0045] The system acquires real-time temperature and smoke concentration at the ignition point to determine whether the fire at the ignition point has been extinguished. If the fire at the ignition point has been extinguished, the system activates the ventilation system corresponding to the target area.

[0046] Specifically, after executing the secondary fire extinguishing procedure, based on the real-time temperature and smoke concentration at the ignition point, it is determined whether the fire at the ignition point has been extinguished. If the fire at the ignition point has been extinguished, the ventilation system corresponding to the target area is activated. This rapidly removes the dense smoke from the archives located in the target area while preventing the removal of inert gases. Furthermore, the entire control process is automatically completed by the ventilation system controller. Based on real-time data from various sensors, the controller adjusts the opening of each ventilation valve using a PID control algorithm and optimizes the balance between smoke extraction and gas retention, enabling the ventilation system to reach the target state in the shortest possible time.

[0047] For example, when the smoke concentration is below 500 ppm and the temperature is below 50°C for 5 consecutive minutes, the fire at the ignition point is considered extinguished. Based on the layout and zoning of the archives, the ventilation system is divided into several independent control units, each covering one or more protected areas. Each unit is equipped with a fresh air valve, an exhaust air valve, and a return air valve. After the fire at the ignition point is extinguished, the fresh air valve and return air valve of the control unit in the fire area are closed, while the exhaust air valve is fully opened to allow the dense smoke to be discharged outdoors through the exhaust duct, with an exhaust rate of no less than 5 times / hour. While the dense smoke is being discharged, other control units maintain normal ventilation to maintain a positive pressure environment within the archives and prevent the spread of dense smoke. The pressure difference between adjacent control units is monitored using differential pressure sensors, and the pressure difference is maintained between 20-50 Pa. Oxygen and inert gas concentration sensors within the fire area monitor the gas concentration in real time. When the oxygen concentration rises above 15% and the inert gas concentration drops below 8%, the fresh air valve is gradually opened and the exhaust air valve is partially closed to refresh the air in the area. Once the inert gas concentration drops below 2%, fully open the fresh air valve and return air valve, close the exhaust valve, and restore normal ventilation to the area, ensuring at least two air changes per hour. Simultaneously, activate the air purification device to filter and purify the incoming fresh air, removing smoke and harmful gases.

[0048] In this embodiment, after the secondary fire extinguishing procedure is executed, if it is determined that the fire at the ignition point has been extinguished, the ventilation system corresponding to the target area is activated to refresh the air in the target area, thereby further ensuring the safety of important materials in the archives.

[0049] In one embodiment, determining whether a potential ignition source exists in the target area based on the acquired initial image data of the target area at the current moment includes:

[0050] The process involves: acquiring initial image data of the target area at the current moment; preprocessing the initial image data using an image processing algorithm to obtain the image data to be tested; extracting image features from the image data to obtain feature vectors, and inputting these feature vectors into a pre-trained feature classification model to obtain multiple types of image features; if the multiple types of image features include fire light wave features, then a potential fire source is identified; the fire light wave features are determined based on historical fire image data.

[0051] Specifically, initial image data of the target area at the current moment is captured using multispectral or infrared sensors. Image processing algorithms include operations such as image denoising, image enhancement, and normalization. The initial image data is preprocessed using these algorithms to obtain the image data to be tested. Image features are extracted from the image data to obtain feature vectors, which include color, texture, and shape. These feature vectors are then input into a pre-trained feature classification model to obtain multiple types of image features. The feature classification model can be a support vector machine or a random forest. If multiple types of image features contain fire light wave characteristics, a potential fire source is identified. Furthermore, the intensity and distribution of the fire light wave characteristics can be analyzed, and combined with environmental factors of the target area, including temperature, humidity, and wind speed, to comprehensively assess the fire hazard and obtain a fire risk level to guide subsequent fire control.

[0052] In this embodiment, the initial image data is processed by an image processing algorithm to extract image features, and the fire wave characteristics in the image are extracted by a pre-trained feature classification model, thereby identifying potential fire sources and improving the accuracy of the fire extinguishing control method of this application.

[0053] In one embodiment, the parameters of the image processing algorithm are adjusted through the following steps:

[0054] Historical image data of historical fire events in the target area and corresponding historical illumination data are acquired. The historical image data and historical illumination data are analyzed to determine the correlation between the image judgment deviation of historical fire events and the illumination change pattern of the target area, so as to train the artificial intelligence model and obtain the correlation model. The current illumination information of the target area is input into the correlation model to obtain the image judgment deviation corresponding to the current illumination information, and the parameters of the image processing algorithm are adjusted according to the image judgment deviation of the current illumination information.

[0055] Specifically, before processing the initial image data using image processing algorithms, the accuracy of identifying potential fire sources varies under different lighting conditions due to the differences between the characteristics of fire wave emission and everyday light reflection. This means there are varying image judgment biases. Therefore, the parameters of the image processing algorithm need to be adjusted based on the image judgment bias under different lighting conditions to accurately extract the fire wave emission characteristics from the image and accurately determine the presence of potential fire sources. First, multispectral sensor data from multiple locations within the target area of ​​the archive are acquired. The data is preprocessed, including noise reduction and smoothing, to obtain time-series data of light intensity. Time-series analysis methods, such as the Autoregressive Moving Average (ARMA) model and seasonal decomposition, are used to model the light intensity data and obtain the light variation patterns over different time periods. Second, by analyzing historical image data of historical fire events and corresponding historical light data, the correlation between image judgment bias of historical fire events and light variations in the target area is determined to establish a correlation model between image prediction bias and light information. Machine learning algorithms such as Support Vector Machine (SVM) and Random Forest are used for training. Finally, the current illumination information of the target region is input into the association model to obtain the image judgment bias corresponding to the current illumination information. The parameters of the image processing algorithm are then adjusted based on the image judgment bias of the current illumination information. The parameters include the threshold and the weights.

[0056] In this embodiment, by establishing a correlation model between image prediction deviation and illumination information, the parameters of the image processing algorithm are adjusted to adapt to fire situation judgment under different illumination conditions, reduce image judgment deviation, and improve the accuracy of the fire extinguishing control method of this application.

[0057] In one embodiment, determining whether an ignition point exists in the target area based on the acquired temperature distribution data of the target area includes:

[0058] Acquire temperature distribution data of the target area and preprocess the temperature distribution data to obtain a temperature distribution map of the target area; calculate the temperature gradient value at each location in the temperature distribution map and determine the location where the temperature gradient value is greater than a preset gradient threshold as the ignition location; if the area formed by multiple consecutive ignition locations is greater than a preset area threshold, then it is determined that there is a fire point in the target area.

[0059] Specifically, when a potential ignition source is detected, a temperature sensor is triggered to begin collecting temperature distribution data of the target area. For example, the data sampling frequency is set to once per second. The data is preprocessed, including noise reduction and smoothing, to obtain a clear temperature distribution map of the target area. The temperature gradient value at each location on the temperature distribution map is calculated, and locations with temperature gradient values ​​greater than a preset gradient threshold are identified as ignition locations. If the area formed by multiple consecutive ignition locations is greater than a preset area threshold, then a fire point is determined to exist in the target area. The preset gradient threshold and preset area threshold are set according to the required detection accuracy.

[0060] In this embodiment, by preprocessing the temperature distribution data of the target area and determining the presence of fire points in the target area based on the obtained temperature gradient map, the occurrence of fire points can be accurately determined, thereby improving the accuracy of the fire extinguishing control method of this application.

[0061] In one embodiment, the step of determining the working fire extinguishing patch and its first spray angle based on the ignition point includes:

[0062] Determine the location of the ignition point; select at least one fire extinguishing patch from among several fire extinguishing patches that is closest to the ignition point as the working fire extinguishing patch; determine the location of the working fire extinguishing patch, and determine the first spray angle based on the location of the working fire extinguishing patch and the location of the ignition point.

[0063] Specifically, the location with the highest temperature gradient value in the temperature gradient map is identified as the ignition point, which is then its two-dimensional coordinates (x, y). Based on these coordinates and the three-dimensional spatial model of the target area, the three-dimensional spatial coordinates (X, Y, Z) of the ignition point are calculated. From a pool of extinguishing pads, at least one pad closest to the ignition point is selected as the working extinguishing pad. Under the default spray intensity, determining the first spray angle based on the positions of the working extinguishing pad and the ignition point can be achieved by inputting the position coordinates into a parabolic theoretical model for calculation, or by using a lookup table to determine the matching spray angle from the known spray range of the working extinguishing pads based on the difference between the position coordinates.

[0064] In this embodiment, the working fire extinguishing patch is determined based on the location coordinates of the ignition point, and the first spray angle is determined based on the location of the working fire extinguishing patch and the location of the ignition point. This allows the fire extinguisher to be activated quickly after the ignition point is determined, improving the fire extinguishing efficiency of the fire control method of this application.

[0065] In one embodiment, the step of acquiring the trajectory of the extinguishing particles ejected at a first spray angle, and determining the second spray angle based on the trajectory, the position of the working extinguishing patch, and the position of the ignition point, includes:

[0066] A high-speed camera captures the motion sequence of fire extinguishing particles sprayed at a first spray angle. Based on the motion sequence, the trajectory is determined, and the position of each frame corresponding to the trajectory is obtained. The expected trajectory of the fire extinguishing particles is determined based on the position of the working fire extinguishing patch and the position of the ignition point. For each frame position, if the difference between the frame position and the corresponding point in the expected trajectory is greater than a preset deviation threshold, the difference is determined as the target difference corresponding to the frame position. All target differences are input into a pre-trained parameter adjustment model to obtain the parameter adjustment value output by the parameter adjustment model. Based on the parameter adjustment value and the first spray angle, the second spray angle is determined.

[0067] Specifically, due to the influence of the spraying accuracy of the fire extinguishing patch and environmental factors, the actual trajectory of the fire extinguishing particles will inevitably deviate from the ideal trajectory. The fire extinguishing particles cannot accurately cover the ignition point. Therefore, it is necessary to dynamically adjust the spraying angle to ensure that the fire extinguishing particles accurately cover the ignition point and minimize damage to surrounding areas. For example, a high-speed camera is placed along the spraying path of the fire extinguishing particles, with a shooting frequency of no less than 500 frames per second and an exposure time of no more than 1 ms. The motion trajectory of the fire extinguishing particles is extracted using the frame difference method to obtain the position of each frame corresponding to the motion trajectory of the fire extinguishing particles. Based on the position of the working fire extinguishing patch and the position of the ignition point, the expected motion trajectory is determined. At time t, the actual coordinates (x, y, z) of the fire extinguishing particles are compared with the theoretical three-dimensional coordinates (x', y', z') of the fire extinguishing particles in the corresponding expected motion trajectory, and the difference Δ = sqrt((x-x')^2 + (y-y')^2 + (z-z')^2). Similarly, for each frame position, the difference is calculated. If the difference between the frame position and the corresponding point in the expected motion trajectory is greater than a preset deviation threshold, it is determined that the trajectory of the extinguishing particles has deviated, and the difference is determined to be the target difference corresponding to the frame position. All target differences are input into a pre-trained parameter adjustment model to obtain the parameter adjustment values ​​output by the parameter adjustment model. The parameter adjustment values ​​and the target differences are positively correlated.

[0068] For example, the parameter adjustment model can be established using a radial basis function neural network. The target difference is used as the input layer, the spray velocity v and spray angle θ as the output layers, and the hidden layers use Gaussian functions. The network is trained using a backpropagation algorithm. The target difference Δ is input into the parameter adjustment model to obtain the adjusted velocity Δv and angle Δθ. Based on the adjusted parameter values ​​and the first spray angle, the second spray angle is determined, i.e., v' = v + Δv, θ' = θ + Δθ. The determined v' and θ' are the second spray angle, thus correcting the spray angle and making the trajectory of the extinguishing particles closer to the expected trajectory.

[0069] In this embodiment, the spray angle of the fire extinguishing patch is adjusted based on the movement trajectory of the fire extinguishing particles to ensure that the fire extinguishing particles accurately cover the ignition point, minimize damage to surrounding archives, and improve the effectiveness of the fire extinguishing control method of this application.

[0070] In one embodiment, determining whether a fire at the ignition point has been extinguished based on the acquired temperature change data and smoke concentration data includes:

[0071] Acquire temperature change data and smoke concentration data at the ignition point; within a preset time period after the working fire extinguishing patch is sprayed, if the current temperature value indicated by the temperature change data is less than the preset temperature threshold, and the current smoke concentration value indicated by the smoke concentration data is less than the preset concentration threshold, then it is determined that the fire at the ignition point has been extinguished.

[0072] Specifically, after the extinguishing particles are sprayed, temperature change data and smoke concentration data at the ignition point are acquired. For example, a thermal imager continuously images the ignition point, acquiring a series of infrared images at a frame rate of at least 10fps. The infrared images are processed to extract the temperature value at the ignition point in each frame, resulting in a temperature time series data T(t), with the temperature value sampling frequency matching the image frame rate. Smoke concentration values ​​are collected in real time using a smoke sensor. After converting the concentration values ​​into digital signals, they are fused with the temperature time series data. Furthermore, temperature change data can be transmitted via a wireless communication module using the MQTT protocol and JSON format; smoke concentration data can be transmitted via an RS485 bus at a transmission rate of at least 9600bps.

[0073] For example, if within one minute after the working fire extinguishing patch is sprayed, the current temperature value indicated by the temperature change data is less than 50°C, and the current smoke concentration value indicated by the smoke concentration data is less than 1000 ppm, then the fire at the ignition point is determined to be extinguished. Otherwise, it is determined that the fire at the ignition point has not been extinguished, and fire extinguishing particles need to be sprayed again.

[0074] In this embodiment, the fire suppression status is judged based on the acquired temperature change data and smoke concentration data to ensure that the fire is effectively controlled, thereby improving the effectiveness of the fire extinguishing control method of this application.

[0075] To better understand the above method, the following details an application example of the fire extinguishing control method based on fire extinguishing stickers in this application.

[0076] First, a high-resolution multispectral camera is installed in the target area. This camera's spectral range covers the visible and near-infrared bands, with a resolution of 4000×3000 pixels. The camera acquires images every 10 seconds, transmitting the data to a server in the monitoring center for analysis. The image data is preprocessed, using a Gaussian filter to remove high-frequency noise, followed by histogram equalization to enhance contrast. Next, the image is normalized to a range of 0-1 for easier feature extraction and classification. In the feature extraction stage, a color histogram and Gabor filter are used to extract color and texture features. The color histogram divides the image into 16 intervals, counting the number of pixels in each interval to form a 16-dimensional feature vector. The Gabor filter uses 5 scales and 8 directions to extract multi-scale and multi-directional texture features, forming a 40-dimensional feature vector. The color and texture features are concatenated to obtain a 56-dimensional feature vector. Support Vector Machines (SVMs) are used to classify the extracted features. The SVM uses the RBF kernel function, and the parameters C and gamma are optimized through a grid search. If the SVM classifies the image as a fire, the area ratio of the fire region in the image is calculated to estimate the fire intensity. Simultaneously, the approximate location of the fire source is estimated based on the location of the fire region in the image. Environmental factors such as temperature, humidity, and wind speed are considered, and a Bayesian network is used to assess the fire hazard. The Bayesian network uses 20 nodes, including image features, environmental factors, and fire hazard, and the conditional probabilities between each node are obtained through training.

[0077] Ten multispectral sensors were installed in the archive, each collecting light intensity data every 5 minutes. The data format was (timestamp, light intensity value), with the light intensity value ranging from 0 to 1000. The collected data underwent noise reduction using a median filter algorithm with a sliding window size of 5 to remove outliers. Then, an exponential smoothing algorithm was used to smooth the data with a smoothing factor of 0.2, resulting in time-series data of light intensity. Next, a seasonal decomposition algorithm was used to decompose the time-series data into trend, seasonal, and random components. The trend component reflects the overall trend of light intensity variation, the seasonal component reflects the periodic variation pattern, and the random component reflects the random fluctuation of light intensity. By analyzing the trend and seasonal components, the patterns of light intensity variation in different time periods were obtained, such as high light intensity during the day and low light intensity at night; gradually increasing light intensity in the morning and gradually decreasing light intensity in the afternoon, etc. From a historical fire event image dataset, 100 images were selected, 50 of which were normal images and 50 were fire images. For each image, its corresponding illumination intensity value is extracted, forming a data pair of (illumination intensity, fire status). The SVM algorithm is used to train the data, employing the RBF kernel function with a penalty coefficient C = 10 and a kernel parameter gamma = 0.1. Through 10-fold cross-validation, the model's accuracy is 90%. The current illumination intensity value is 500; inputting it into the SVM model yields an image judgment bias of 0.2. Based on this bias, the threshold parameter of the image processing algorithm is adjusted; the original threshold is 100, and the adjusted threshold is 100 + 0.2 * 100 = 120. The currently acquired image data is input into the adjusted image processing algorithm. The algorithm first preprocesses the image, including grayscale conversion and denoising. Then, it extracts the image's color histogram features and LBP texture features, concatenating these two features into a feature vector. The SVM algorithm is used to classify the feature vector to determine if the image contains a fire. If the classification result indicates a fire, an alarm is triggered, and the alarm information is sent to the fire department and archive management personnel. If the classification result is normal, the monitoring of image data at the next moment continues. Through the above steps, a real-time fire monitoring and early warning system for archives based on multispectral sensor data and image processing technology has been implemented. The system can dynamically adjust the parameters of the image processing algorithm according to changes in lighting conditions, improving the accuracy and reliability of fire assessment.

[0078] Second, 100 temperature sensors were installed in a warehouse, arranged in a 10×10 grid with a row and column spacing of 2 meters. When an open flame appeared in a part of the warehouse, the nearby temperature sensors detected a sudden temperature rise, triggering a fire alarm. The system began collecting data from all temperature sensors at a sampling frequency of once per second for 10 seconds. The collected temperature data was then denoised using a median filter algorithm with a sliding window size of 3 to remove outlier temperature values. A moving average algorithm was then used to smooth the data with a smoothing window size of 5, resulting in a smoothed temperature distribution map. Next, temperature gradient analysis was performed on the temperature distribution map. First, the temperature gradient value at each location was calculated using the formula: G(x,y)=sqrt{(frac{T(x+1,y)-T(x-1,y)}{2})^2+(frac{T(x,y+1)-T(x,y-1)}{2})^2}, where T(x,y) represents the temperature value at location (x,y). Then, the presence of a fire is determined based on the magnitude of the temperature gradient. If the temperature gradient at a certain location exceeds a preset threshold (e.g., 50℃ / m), and at least 5 of the surrounding 8 locations also exceed this threshold, a fire is identified at that location. By analyzing the distribution of the temperature gradient values, the extent of the fire zone is determined. Within the fire zone, the location with the largest temperature gradient is identified as the ignition point. The two-dimensional coordinates of the ignition point are calculated based on its row and column number in the temperature sensor grid. For example, if the ignition point is located in row 3 and column 6, its two-dimensional coordinates are (5, 11), since the row and column spacing is 2 meters. Finally, the three-dimensional coordinates of the ignition point are calculated based on its two-dimensional coordinates and the three-dimensional spatial model of the warehouse. Assuming the warehouse height is 5 meters and the ignition point is located on a shelf 2 meters above the ground, its three-dimensional coordinates are (5, 11, 2). This coordinate information is encapsulated into a JSON message and sent to the fire extinguishing patch control system via the MQTT protocol. After the control system analyzes the coordinates of the fire point, it controls the nearest fire extinguishing patch to aim at that location and spray extinguishing agent to put out the fire. The entire fire detection and extinguishing process is completed automatically without human intervention, greatly improving the efficiency and safety of fire response.

[0079] The fire extinguishing patch control system receives a JSON message via a ZigBee wireless communication module. The message content is {"X":5.0,"Y":11.0,"Z":2.0}, representing the three-dimensional spatial coordinates of the ignition point as (5.0, 11.0, 2.0). After parsing the coordinate values, the system queries the fire extinguishing patch location information database, which stores the ID number and corresponding three-dimensional coordinates of each fire extinguishing patch. By calculating the Euclidean distance between the ignition point coordinates and the coordinates of each fire extinguishing patch, the three fire extinguishing patches with the smallest distances are identified. Their ID numbers are A-10, B-25, and C-07, and their coordinates are (3.5, 10.0, 2.5), (6.2, 12.1, 1.8), and (4.7, 9.8, 2.2), respectively. The control system sends a start command to these three fire extinguishing patches, which includes the ignition point coordinate information. Upon receiving the command, the built-in microcontroller control unit of the fire extinguishing patch begins to calculate the elevation and azimuth angles of the fire extinguishing particle launching device. Taking fire extinguisher patch A-10 as an example, the coordinates of the ignition point are (5.0, 11.0, 2.0), and the coordinates of the fire extinguisher patch are (3.5, 10.0, 2.5). Then the horizontal distance is D = sqrt{(5.0-3.5)^2 + (11.0-10.0)^2} = 1.80 meters, and the vertical height difference is Zz = 2.0-2.5 = -0.5 meters. Therefore, the pitch angle is theta = arctan(-0.5 / 1.80) = -15.52 π / c, and the azimuth angle is varph i = arctan(11.0-10.0 / 5.0-3.5) = 53.13 π / c. The motor drive system of the fire extinguishing particle launcher adjusts the pitch angle of the launch tube to -15.52° and the azimuth angle to 53.13° based on the angle values ​​calculated by the control unit, aligning the launch tube with the coordinates of the fire point. After the angle adjustment is complete, the control unit triggers the fire extinguishing particle launcher, spraying 500 grams of fire extinguishing particles along the launch tube with an initial velocity of 20 m / s. During flight, the fire extinguishing particles are subject to gravity, forming a parabolic trajectory. After 0.5 seconds, the fire extinguishing particles land near the fire point, with a vertical distance of approximately 1.2 meters and a horizontal distance of approximately 0.8 meters from the ground, within an acceptable range of deviation from the fire point. Upon landing, the outer shell of the fire extinguishing particle breaks under the impact force, releasing the internal extinguishing agent, which rapidly diffuses to the fire source. There, it reacts chemically with the open flame, producing large amounts of carbon dioxide and water vapor, isolating the fire source from external oxygen, thereby achieving the purpose of extinguishing the fire. At the same time, two other fire extinguishing patches were also used to extinguish the fire. The fire extinguishing particles from multiple directions focused on covering the fire source, ensuring the effectiveness of the firefighting.The entire fire extinguishing process is accomplished through the coordinated operation of multiple fire extinguishing patches. The fire extinguishing patch control system automatically dispatches the nearest fire extinguishing patch based on the coordinates of the fire point and controls it to strike precisely. The entire process requires no manual intervention, has a fast response speed, and high positioning accuracy, and can extinguish the fire in its early stages, preventing the fire from spreading and expanding, and greatly reducing fire losses.

[0080] Third, the system captures images of the fire extinguishing stickers using cameras to determine if they have been activated. If not, the stickers are activated. High-definition cameras are positioned near the fire extinguishing stickers, with their field of view covering the entire sticker. The cameras are connected to the fire extinguishing sticker control system via wired or wireless means. The cameras continuously capture images of the fire extinguishing stickers at a frequency of 5 frames per second, transmitting the image data to the control system in real time. The control system uses image processing algorithms to analyze the received images. The algorithms preprocess the images, including grayscale conversion, noise reduction, and enhancement, to obtain images suitable for analysis. A template matching algorithm is then used to identify the fire extinguishing stickers. The template used is a frontal image of the fire extinguishing sticker. The template is calculated... The system uses image similarity to determine if a fire extinguishing sticker exists in the image. If a fire extinguishing sticker is present, its status is further assessed, characterized by its appearance features such as indicator light color and emitter angle. Based on a pre-established fire extinguishing sticker status feature database, the features of the current fire extinguishing sticker are extracted and compared. If the features match the inactive state (indicator light is red or emitter angle is 0 degrees), the fire extinguishing sticker is determined to be inactive. If an inactive fire extinguishing sticker is detected and a fire exists, the control system sends an activation command to the fire extinguishing sticker. The command is transmitted via wired or wireless means. Upon receiving the command, the fire extinguishing sticker executes the activation operation until its status features match the activation state.

[0081] A high-definition camera with a 60-degree field of view and a resolution of 1920×1080 is installed 1 meter directly in front of the fire extinguisher patch, completely covering the entire image of the patch. The camera is connected to the industrial control computer of the fire extinguisher patch control system via a USB 3.0 interface, with a transmission bandwidth of 5Gbps, sufficient for high-speed image transmission. The camera continuously captures images of the fire extinguisher patch at a frequency of 5 frames per second, with each frame being 2MB in size. The image data is transmitted to the industrial control computer's memory in real time as a stream, with a transmission latency of less than 50 milliseconds. The industrial control computer runs an image processing algorithm to analyze each frame in real time. First, the image is preprocessed, using a median filter algorithm to remove salt-and-pepper noise from the image, with a filter window size of 5×5. Then, a histogram equalization algorithm is used to enhance the image contrast and improve the visual effect. Finally, the image is converted to grayscale to reduce the computational load of subsequent processing. The preprocessed image is used for fire extinguisher patch recognition, employing a template matching method based on SURF (Accelerated Robust Features). First, SURF feature points are extracted from the template image, which is a high-resolution frontal image of the fire extinguisher sticker. The number of feature points is set to 500. Then, SURF feature points are extracted from the image to be identified. Two sets of feature points are matched to obtain the number of matching point pairs n. The matching degree S = n / 500 is calculated. If S is greater than 0.8, the fire extinguisher sticker is determined to exist in the image; otherwise, it is determined to be background. When the fire extinguisher sticker is detected in the image, its state is further determined. The fire extinguisher sticker has three states: standby, start-up, and warning, corresponding to green, yellow, and red indicator lights, respectively. The HSV color space is used to classify the pixels in the indicator light area. The number of pixels with H values ​​in the range of [35,77], [78,99], [0,10]∪[156,180] is counted, corresponding to green, yellow, and red, respectively. The color with the most pixels is selected as the indicator light state. At the same time, the angle of the fire extinguisher sticker's emitter tube is detected. With vertical upward as 0 degrees and clockwise as the positive direction, the Hough linear transform is used to detect the edge of the emitter tube and calculate its angle with the vertical line with an accuracy of 1 degree. The status of the fire extinguishing patch is determined by the indicator light color and the angle of the emitter tube. A green indicator light and an emitter tube angle of 0 degrees indicate standby mode; a yellow indicator light and an emitter tube angle greater than 0 degrees indicate activation mode; and a red indicator light and an emitter tube angle greater than 0 degrees indicate alert mode. When the system detects a fire, the control system sends an activation command to the fire extinguishing patch in standby mode. The command is transmitted via the CAN bus, and the data frame format is: frame header (1 byte) + ID number (4 bytes) + command type (1 byte) + data length (2 bytes) + data content (n bytes) + frame tail (2 bytes). The command type is 0x01, indicating an activation command, and the data content is the coordinates of the ignition point (X, Y, Z). After receiving the activation command, the control unit parses the ignition point coordinates, inputs them into the angle calculation model to obtain the target angle of the emitter tube, and then controls the motor to drive the emitter tube to rotate. Simultaneously, the yellow indicator light illuminates, indicating that the system has entered the activation state.When the launch tube reaches the target angle, the control unit triggers the fire extinguishing projectile launching device. Under pressure, the fire extinguishing projectile is ejected at high speed along the launch tube and flies towards the fire point to begin the fire extinguishing operation.

[0082] Fourth, a high-speed camera is placed every 10 meters along the flight path of the extinguishing particles. The camera captures at a frequency of 1000 frames per second, with an exposure time of 0.5 ms and a resolution of 1280×720. During one fire extinguishing operation, the extinguishing patch sprays 50 extinguishing particles, each with a diameter of 2 cm, an initial velocity of 30 m / s, and a spray angle of 60°. Based on the flight time of the extinguishing particles and the positions of the cameras, the time when the extinguishing particles pass through each camera is estimated. When the extinguishing particles enter the field of view of the cameras, the image processing unit uses the frame difference method to extract the motion trajectory of the extinguishing particles. The specific algorithm is as follows: subtract the grayscale image of the current frame from the grayscale image of the previous frame to obtain a difference image; binarize the difference image to obtain a binary image of the extinguishing particles; perform morphological processing on the binary image to remove noise and holes; calculate the centroid coordinates (x, y) of the binary image, which is the position of the extinguishing particle in the current frame. By tracking the position of the extinguishing particles in each frame, their motion trajectory is obtained. Based on the position of the fire extinguishing patch (0, 0, 0) and the spray angle θ = 60°, a theoretical motion model of the fire extinguishing particles is established. Assuming the fire extinguishing particles move in the xy plane, neglecting displacement along the z-axis, the theoretical coordinates of the fire extinguishing particles at time t are: x' = v0cos(θ) ty' = v0sin(θ)t - 0.5gt^2, where v0 is the initial velocity of 30 m / s and g is the gravitational acceleration of 9.8 m / s^2. The actual coordinates (x, y) of the fire extinguishing particles captured by the camera are compared with the theoretical coordinates (x', y'), and the deviation value Δ = sqrt((x - x')^2 + (y - y')^2) is calculated. Assuming that at t = 0.5s, the actual coordinates are (13, 10) and the theoretical coordinates are (15, 11), then the deviation value Δ = sqrt((13 - 15)^2 + (10 - 11)^2) = 2.24 meters. A deviation threshold of 1 meter is set. When Δ > 1, it is determined that the trajectory of the extinguishing particles has deviated, and the spray parameters need to be adjusted. An RBF neural network is used to establish the mapping relationship between the deviation value and the spray parameters. The network structure is as follows: one neuron in the input layer receives the deviation value Δ; ten neurons in the hidden layer activate the Gaussian function exp(-||xc||^2 / (2*σ^2)), where c is the data center and σ is the variance; two neurons in the output layer output the speed adjustment Δv and the angle adjustment Δθ, respectively. Assuming the deviation value Δ = 2.24 at t = 0.5s, it is input into the RBF network, resulting in the outputs Δv = -1.5m / s and Δθ = -3°. The adjusted speed v' = 30 - 1.5 = 28.5m / s, and the adjusted angle θ' = 60 - 3 = 57°. v' and θ' are sent to the extinguishing patch to control the launching device to readjust the speed and angle, making the trajectory of the next extinguishing particle more accurate. Continue to track the trajectory of the extinguishing particles and constantly adjust the spray parameters until the landing point of the extinguishing particles is within the target area.By monitoring in real time and making dynamic adjustments, the effectiveness of firefighting was ensured while avoiding accidental damage to non-target areas.

[0083] Obtain the deviation value Δ between the actual position (x, y, z) and the theoretical position (x', y', z') of the extinguishing particles calculated in the previous step. Δ reflects the degree of deviation between the trajectory of the extinguishing particles and the expected trajectory. Based on the deviation value Δ, calculate the required adjustment spray angle value Δθ and spray intensity value Δv by looking up a table or interpolation. The adjustment value is proportional to the deviation value; the larger the deviation, the larger the adjustment range. Input Δθ and Δv into the control unit of the rotating component. The control unit outputs a motor control signal to drive the motor to rotate the extinguishing particle launching device around the horizontal and vertical directions, adjusting the spray angle. The formula for calculating the rotation angle is: θ' = θ + Δθ, where θ is the current angle and θ' is the adjusted target angle. Δθ is the angle adjustment value; the rotation process uses a PID control algorithm to dynamically adjust the motor speed and acceleration based on the angle feedback signal to ensure that the target angle is reached quickly and smoothly; the spray intensity is adjusted by changing the pressure in the pressure chamber, and the pressure is proportional to the initial velocity of the extinguishing particles; the gas flow rate entering the pressure chamber is adjusted by controlling the opening of the air valve, thereby adjusting the pressure; the pressure calculation formula is: P'=P+ΔP, where P is the current pressure, P' is the adjusted target pressure, and ΔP is the pressure adjustment value; the value range of ΔP is [-0.2, 0.2] MPa, and the adjustment step is 0.05 MPa; the opening of the air valve is controlled by the PLC to achieve rapid and precise pressure adjustment.

[0084] Assuming that at t = 1s, the actual position of the extinguishing particles is (15, 8, 3), and the theoretical position is (18, 10, 5), then the deviation value Δ = sqrt((15-18)^2 + (8-10)^2 + (3-5)^2) = 4.36. Looking up the deviation value in the table, the angle adjustment value Δθ = (-2°, -1°) and the spray intensity adjustment value Δv = 0.1MPa are obtained. Δθ is input to the control unit of the rotating component, and the control unit outputs a PWM signal to drive the stepper motor to rotate. The motor's rotation angle θ' = θ + Δθ. Assuming the current horizontal angle θh = 45° and the vertical angle θv = 30°, then the adjusted target angles θh' = 45 - 2 = 43° and θv' = 30 - 1 = 29°. The control unit uses an incremental PID algorithm to calculate the motor control quantity based on the angle feedback signal. The formula for calculating the control quantity is: ΔU=Kp×(e(k)-e(k-1))+Ki×e(k)+Kd×(e(k)-2e(k-1)+e(k-2)), where Kp=2, Ki=0.5, Kd=0.1, e(k) is the angle error of the k-th sampling, and ΔU is the control increment. The motor speed v is proportional to the control increment, v=Kv×ΔU, where Kv=10, and the unit of speed is r / s. The motor acceleration a=Ka×(v-v0), where Ka=20, v0 is the speed of the previous moment, and the unit of acceleration is r / s^2. By controlling the motor speed and acceleration, the horizontal angle is adjusted to 43° and the vertical angle is adjusted to 29° within 1 second, with the angle error controlled within ±0.1°. The pressure control unit inputs Δv = 0.1 MPa, and outputs a 4-20 mA analog signal to control the opening of the proportional valve. The relationship between the valve opening y and the control current I is: y = Ky × (I - 4), where Ky = 6.25, and the unit of opening is %. Assuming the current pressure Pc = 1.0 MPa, the adjusted target pressure P' = 1.0 + 0.1 = 1.1 MPa. The pressure control unit uses a positional PID algorithm to calculate the valve control quantity based on the pressure feedback signal. The formula for calculating the control quantity is: I = Kp × (SP - PV) + Ki × ∑(SP - PV) × Δt + Kd × (PV - PV0) / Δt, where Kp = 2, Ki = 0.2, Kd = 0.1, SP is the pressure setpoint, PV is the pressure feedback value, Δt is the sampling period, and PV0 is the pressure feedback value of the previous period. The valve opening adjustment step Δy = Ky × (I - I0), where I0 is the control current at the previous moment. By adjusting the valve opening, the pressure is adjusted to 1.1 MPa within 0.5 s, with the pressure error controlled within ±0.01 MPa. After the angle and pressure are adjusted, the control unit of the fire extinguishing patch triggers the ignition command of the fire extinguishing device, igniting the fire pack. The high-pressure gas generated by the combustion of the fire pack propels the fire extinguishing bullet out at high speed along the launch tube. The initial velocity of the fire extinguishing bullet is v0 = 30 + Kp × Δv = 35 m / s, where Kp = 50.During its flight, the fire extinguishing bomb is affected by gravity and air resistance, and its trajectory is a parabola. The coordinates of the landing point are (x, y, z) = (18, 10, 0). The distance error between the landing point and the target point (18, 10, 5) is Δd = 5m, which meets the accuracy requirements.

[0085] Fifth, after the extinguishing particles are sprayed, a thermal imager installed above the fire area begins continuous imaging of the fire point. The thermal imager has a resolution of 640×480, a frame rate of 25fps, a field of view of 60°, and a working wavelength of 8-14μm. The thermal imager acquires one frame of infrared image every 40ms, and the image data is transmitted to the image processing unit in real time via a USB 3.0 interface. The image processing unit uses an FPGA chip for parallel computing. First, it performs non-uniformity correction on the infrared image using a two-point correction method, acquiring reference images of the sensor at low temperature (20℃) and high temperature (120℃) respectively. The corrected image Ic is calculated according to the formula Ic=(I-Il) / (IhIl)×(Th-Tl)+Tl, where I is the original image, Il and Ih are the low-temperature and high-temperature reference images respectively, and Tl and Th are the low-temperature and high-temperature temperature values ​​respectively. Then, the corrected image is filtered and denoised using a 5×5 median filter to remove salt-and-pepper noise from the image. Finally, the region with the highest temperature in the image is extracted, the average temperature value of that region is calculated, and the temperature measurement value of the current frame is obtained.

[0086] 100 frames of infrared images were continuously acquired to obtain a temperature time series T(t), where t ranges from [0, 4] s and the time interval is 40 ms. T(t) was sent to the fire extinguishing patch control system via the MQTT protocol. The data frame structure was {timestamp, temperature value}, with units of ms and ℃ respectively, and each data frame was 8 bytes in size. Simultaneously, smoke sensors installed within the fire extinguishing area collected smoke concentration values ​​in real time. The sensor's range was 0-10000 ppm, its resolution was 10 ppm, and its response time was 1 s. The sensor collected the concentration value every 100 ms and converted it into a digital signal via a 12-bit ADC. The digital signal was transmitted to the fire extinguishing patch control system via an RS485 bus at a transmission rate of 115200 bps. After receiving the temperature and smoke concentration data, the fire extinguishing patch control system first synchronized the data by aligning the timestamps of the two sets of data to obtain a synchronized data pair (T, S). The data is then smoothed using an exponentially weighted average method. The smoothed temperature value Ts(t) = αT(t) + (1-α)Ts(t-1), and the smoothed concentration value Ss(t) = αS(t) + (1-α)Ss(t-1), where α is the smoothing coefficient with a value of 0.2. Based on the smoothed data pair (Ts, Ss), a fuzzy control algorithm is used to determine whether the fire is under control. First, the temperature and concentration values ​​are fuzzified. The fuzzy set for temperature is {low, medium, high}, and the fuzzy set for concentration is {low, medium, high}. After fuzzification, the temperature membership vector Td and the concentration membership vector Sd are obtained. Then, fuzzy inference is performed based on the fuzzy rule base, which contains 9 rules, such as: IF Td = high AND Sd = high THEN fire = uncontrolled. The inference result yields the fire membership vector Fd. Finally, Fd is defuzzified to obtain the fire control level value Fc. The value of Fc ranges from [0, 1], with a larger value indicating better fire control. If Fc is greater than 0.9 and remains stable within 1 minute, the fire is considered to be under effective control, and the spraying of extinguishing particles is stopped. If Fc is less than 0.7 and shows a downward trend, the fire is considered to be worsening, and the next round of spraying is initiated. If Fc fluctuates between 0.7 and 0.9, the fire is considered to be basically under control, but continued monitoring is still required.

[0087] If the fire is determined to be not fully under control, an infrared temperature sensor is activated to monitor the temperature at the ignition point. If the detected ignition point temperature exceeds a preset threshold, a secondary fire suppression program is automatically initiated. This program includes using adjacent fire extinguishing pads or activating an inert gas generator in the corresponding fire area to release inert gas. Based on the fire control level value Fc, if Fc remains below 0.7 for 5 consecutive minutes, the fire is considered not fully under control, and the secondary fire suppression mode is activated. In this mode, an infrared temperature sensor installed near the ignition point is activated. The sensor has a temperature range of 0-1000℃, an accuracy of 1℃, and a response time of 100ms. The sensor collects the ignition point temperature every 500ms and compares it to a preset threshold Tc, which has a default value of 200℃ and is adjusted based on environmental conditions and material composition. If three consecutive temperature readings exceed Tc, a risk of reignition is identified, and the secondary fire suppression program is activated. Based on the location of the fire, the program selects the 1-2 nearest unactivated fire extinguishing pads and sends activation commands to them. Upon receiving the activation command, the fire extinguishing pads begin extinguishing the fire, following the same process as the main program, including attitude adjustment and fire extinguishing grenade launch, until the temperature at the fire point drops below Tc. If the secondary extinguishing program still fails to effectively control the fire, the inert gas extinguishing system is activated. Based on the location information of the fire point, the corresponding protected area is determined, and an activation command is sent to the inert gas generator in the corresponding area. After the inert gas generator is activated, the air supply to the protected area is first cut off, and nitrogen or carbon dioxide inert gas is released into the area. When the gas concentration reaches 13% or higher, the oxygen concentration in the area is reduced to below 12%, preventing the combustibles from continuing to burn. The inert gas continues to be released within the protected area for 3-5 minutes until the infrared camera can no longer detect open flames, the smoke concentration is below 500 ppm, and the fire point temperature is below 80°C. At this point, the fire is considered completely extinguished, the gas release is stopped, and ventilation in the area is restored.

[0088] Specifically, after the main fire suppression procedure ends, the fire suppression patch control system continuously monitors the fire control level value Fc. If Fc remains below 0.7 for 5 minutes and shows a downward trend, it is determined that the fire is not fully under control, and the secondary fire suppression mode is activated. For example, if Fc = 0.75 at t = 5 minutes, Fc = 0.68 at t = 7 minutes, and Fc = 0.65 at t = 10 minutes, then the activation conditions for the secondary fire suppression mode are met. In the secondary fire suppression mode, infrared temperature sensors located within a range of (15, 22, 82) meters from the ignition point begin to operate. The sensors use uncooled focal plane detectors with a sensitivity of 50 mK and an output signal of 4-20 mA. The sensors collect temperature values ​​every 500 ms and transmit them to the fire suppression patch control system via an RS485 bus. If three consecutive temperature values ​​of {210, 225, 235}℃ are collected, all exceeding the preset threshold Tc = 200℃, it is determined that reignition may occur at the ignition point, requiring the initiation of a secondary fire suppression procedure. Based on the ignition point coordinates, backup fire suppression patches are searched within a 1.5-meter radius, resulting in two closest patches, numbered A15 and B27, with coordinates (14, 24, 9) and (13, 21, 7.5), respectively. The control system sends activation commands to both patches in parallel, including parameters such as the ignition point coordinates, extinguishing agent type, and spray intensity. Upon receiving the commands, A15 and B27 adjust their attitude, aligning their launch tubes with the ignition point, and then simultaneously initiate the fire suppression projectile launch procedure. The two patches alternately launch fire suppression projectiles at a rate of 2 times / second, with 5 projectiles launched per burst at a speed of 50 m / s. The fire extinguishing process lasts for 30 seconds. During this period, the infrared camera continuously monitors the temperature change at the ignition point. If the temperature drops below Tc and the smoke concentration is below 1000ppm, it is determined that the risk of reignition has been eliminated, and the fire extinguishing projectiles are stopped.

[0089] If, after the secondary fire suppression procedure is completed, the ignition point temperature remains above Tc and shows a clear upward trend, the inert gas fire suppression system is activated. The system determines the protected area A2 based on the ignition point coordinates (15, 22, 8) and sends a start command to the inert gas generator in area A2. The generator first closes the fresh air valve and exhaust valve in area A2 to isolate airflow, then opens the inert gas storage tank valve, releasing high-purity nitrogen or carbon dioxide into area A2 at a rate of 1000 L / min. During the gas release, oxygen concentration sensors installed in the area monitor changes in oxygen concentration in real time. When the concentration drops below 12%, the ignition point of the combustible material increases significantly, greatly reducing the probability of continued combustion. The gas release lasts for 5 minutes, during which an infrared camera captures an image of area A2 every 10 seconds for flame and smoke detection. If no flames or smoke are detected in the image for one minute, and the temperature at the ignition point drops below 80°C, the fire is considered to be completely extinguished. The gas valves are then shut off, ventilation in the area is gradually restored, and a fire clearance signal is sent to the fire department.

[0090] Sixth, after the fire extinguishing process, the smoke sensors in the A2 protected area collect smoke concentration values ​​every 10 seconds, and the temperature sensors collect temperature values ​​every 30 seconds. If 300 consecutive smoke concentration values ​​are below 500 ppm and 100 consecutive temperature values ​​are below 50°C, it is determined that the fire in the A2 area has been completely extinguished. The archive's ventilation system is divided into 10 independent control units, numbered from VA01 to VA10, with unit VA02 covering the A2 protected area. Before the fire, the fresh air valve in unit VA02 was open at 80%, the exhaust valve at 20%, and the return air valve was fully open, maintaining a positive pressure state in the area. After the fire in the A2 area was extinguished, the control strategy of unit VA02 changed accordingly: the fresh air valve and return air valve were closed, the exhaust valve was fully opened, and the exhaust fan frequency was increased to 50Hz, so that the exhaust volume reached 12000 m³ / h. 3 / h, equivalent to 6 air changes per hour. Meanwhile, units VA01 and VA03 adjacent to area A2 maintain normal ventilation, with the fresh air valve opening maintained at 60%, the exhaust valve at 40%, and the return air valve at 50%, creating a slightly positive pressure environment with a pressure difference of approximately 30 Pa in both areas. During the smoke extraction process, oxygen and carbon dioxide concentration sensors installed in area A2 continuously monitor changes in gas concentration. Initially, the oxygen concentration in the area is 12%, and the carbon dioxide concentration is 15%. As the smoke extraction progresses, the oxygen concentration gradually increases, while the carbon dioxide concentration gradually decreases. When the oxygen concentration rises to 15%, the fresh air valve begins to open gradually, increasing the valve opening at a rate of 10% per minute until it reaches 50%; simultaneously, the exhaust valve opening is decreased at a rate of 10% per minute until it drops to 50%. During this process, the carbon dioxide concentration in the area further decreases. When the concentration falls below 8%, the fresh air valve fully opens, the exhaust valve opening decreases to 20%, and the return air valve fully opens, restoring normal ventilation to the area. The entire smoke extraction process is automatically implemented by the ventilation controller of unit VA02. The controller uses a PLC as its core, dynamically adjusting the opening of the fresh air valve, exhaust air valve, and return air valve based on real-time data collected by various sensors using a PID control algorithm. It also adjusts the frequency of the exhaust fan in real-time according to the area pressure difference. The PID controller has a proportional gain Kp = 2, an integral gain Ki = 0.5, a derivative gain Kd = 0.1, and a sampling period Ts = 5s. The valve opening adjustment formula is: u(k) = Kp × e(k) + Ki × ∑e(k) × Ts + Kd × / Ts, where u(k) is the valve opening in the k-th sampling period, and e(k) is the oxygen concentration deviation or pressure difference deviation in the k-th sampling period. Through the controller's optimized calculations, unit VA02 completes the smoke extraction and ventilation restoration tasks within 30 minutes, ensuring the air quality and environmental safety of the archives.

[0091] Seventh, the system acquires activation information for fire extinguishing patches through a dynamic monitoring system, integrates information on activated fire extinguishing patches, and regularly generates fire extinguishing patch reports to remind relevant personnel to replace activated fire extinguishing patches. A status sensor is installed on each fire extinguishing patch, and the status sensor connects to the dynamic monitoring system via a wireless communication module to report the status information of the fire extinguishing patch in real time, including the fire extinguishing patch ID, location coordinates, status code, and battery level. The dynamic monitoring system adopts an Internet of Things (IoT) architecture, including a data acquisition layer, a data transmission layer, and a data application layer. The data acquisition layer consists of distributed status sensors; the data transmission layer uses LoRa and NB-IoT low-power wide-area network technologies to upload the collected status data to a cloud server; the data application layer uses big data analytics to process and mine the status data. After receiving the status data, the cloud server performs data cleaning and preprocessing, removing redundant and erroneous data. The data is categorized according to the fire extinguishing patch ID and stored in corresponding data tables, with one record for each fire extinguishing patch. For each fire extinguishing patch record, its status code field is extracted, and the status code is used to determine whether the fire extinguishing patch has been activated. The status code uses 2-byte encoding. The first 8 bits indicate the type of fire extinguishing patch, and the last 8 bits indicate the specific status. 0x01 indicates standby status, 0x02 indicates active status, and 0x03 indicates fault status. Fire extinguishing patch records with status code 0x02 are extracted to generate an active fire extinguishing patch list, which includes the patch ID, location coordinates, and activation time. The activation time is calculated using the timestamp of the status code change. A monthly report is generated, summarizing the active fire extinguishing patch list for the previous month. The report is in PDF format and includes the number of active fire extinguishing patches, a heat map of their distribution, and status analysis. It is sent to relevant maintenance personnel via email or SMS. Maintenance personnel replace the active fire extinguishing patches based on the report. During replacement, the new fire extinguishing patch information, including the patch ID, location coordinates, and initial status, needs to be entered into the dynamic monitoring system to ensure the continuity and accuracy of the monitoring data. The dynamic monitoring system is regularly maintained and upgraded, including hardware inspection, software updates, and database backups, to improve system reliability and stability.

[0092] Specifically, 500 fire extinguishing stickers were deployed within the archives. Each sticker was equipped with a status sensor, which used a ZigBee wireless communication module operating at 2.4GHz with a transmission power of 10mW and a transmission distance of up to 100 meters. The status sensor collected status data from the fire extinguishing stickers every 10 seconds, including the sticker ID, location coordinates (x, y, z), status code, and battery voltage, with a data length of 16 bytes. This status data was transmitted via the ZigBee network to a data aggregation node, and then uploaded to a cloud server via an NB-IoT gateway. The NB-IoT gateway supports up to 50,000 nodes, with an uplink speed of up to 60kbps and a coverage radius of up to 10 kilometers. The cloud server used the Alibaba Cloud IoT platform, providing device management, data storage, and a rule engine. Upon receiving the status data, the server first performed format verification and deduplication, and then unpacked and stored the data according to the fire extinguishing sticker ID. The storage uses a MySQL database. The server scans the data table every 10 minutes via a scheduled task, extracting records with a status_code field of 0x02 and generating a list of activated patch messages. The timestamp in the list is calculated by comparing the current record's update_time field with the previous record's, i.e., start_time = update_time - LAG(update_time)OVER(PARTITIONBYpatch_idORDERBYupdate_time). The monthly report task is implemented using the Quartz scheduling framework and is triggered at 2:00 AM on the 1st of each month. The report generation process is as follows: First, count the number of activated fire extinguishing patches from the previous month using the formula COUNT(DISTINCTpatch_id). Second, analyze the status distribution of activated fire extinguishing patches, calculating the number and percentage of patches in normal, faulty, and low-battery states. Third, generate a PDF report document by using the iText library to insert the statistical results and charts into a PDF template, producing the final report file. Fourth, send the report via email by using the JavaMailing library to attach the PDF file to a pre-defined list of maintenance personnel's email addresses. Maintenance personnel will replace activated or faulty fire extinguishing patches according to the report content. During replacement, the information of the new fire extinguishing patch is entered into the management system by scanning the QR code on the patch, including the patch ID, location coordinates, and initial status. After successful entry, the old fire extinguishing patch will be automatically marked as inactive. The system undergoes preventative maintenance quarterly, inspecting and cleaning equipment such as data collection nodes, gateways, and servers, and upgrading firmware when necessary. Simultaneously, parameters such as the deployment density and spray angle of the fire extinguishing patches are optimized according to the latest fire protection codes and technical standards to improve fire extinguishing efficiency.Maintenance records are stored in the operation and maintenance management system in the form of work orders, including information such as equipment name, inspection items, optimization parameters, and the person in charge, for subsequent traceability and auditing.

[0093] See Figure 2 The diagram shows a structural schematic of a fire extinguishing control system based on fire extinguishing patches provided in an embodiment of this application. The system includes:

[0094] The potential ignition source determination module 201 is used to determine whether a potential ignition source exists in the target area based on the initial image data of the target area at the current moment.

[0095] The fire extinguishing patch determination module 202 is used to determine whether there is an ignition point in the target area based on the temperature distribution data of the target area if a potential fire source is found. When there is an ignition point, the working fire extinguishing patch and its first spray angle are determined based on the ignition point. The working fire extinguishing patch is at least one of a number of fire extinguishing patches set in the target area.

[0096] The fire extinguishing patch driving module 203 is used to drive the working fire extinguishing patch to spray fire extinguishing particles toward the fire point at a first spray angle.

[0097] The spray angle adjustment module 204 is used to acquire the movement trajectory of the fire extinguishing particles sprayed at the first spray angle, and determine the second spray angle based on the movement trajectory, the position of the working fire extinguishing patch and the position of the ignition point, and drive the working fire extinguishing patch to spray fire extinguishing particles toward the ignition point at the second spray angle.

[0098] In one embodiment, the system further includes a secondary fire extinguishing module, used to determine whether the fire at the ignition point has been extinguished based on the acquired temperature change data and smoke concentration data at the ignition point; if the fire at the ignition point has not been extinguished, and the current temperature value of the ignition point indicated by the temperature change data is greater than a preset temperature, then at least one fire extinguishing patch adjacent to the working fire extinguishing patch among a plurality of fire extinguishing patches is driven to spray fire extinguishing particles toward the ignition point, or the output end of the inert gas generator corresponding to the ignition point is driven to release inert gas toward the ignition point.

[0099] In one embodiment, the system further includes a gas management module for acquiring the real-time temperature value and real-time smoke concentration at the ignition point to determine whether the fire at the ignition point has been extinguished. If the fire at the ignition point has been extinguished, the ventilation system corresponding to the target area is activated.

[0100] In one embodiment, the potential fire source determination module 201 is further configured to acquire initial image data of the target area at the current moment; perform image preprocessing on the initial image data using an image processing algorithm to obtain image data to be tested; extract image features from the image data to be tested to obtain feature vectors of the image data to be tested, and input the feature vectors into a pre-trained feature classification model to obtain multiple types of image features; if the multiple types of image features include fire light wave features, then a potential fire source is determined; wherein, the fire light wave features are determined based on historical fire image data.

[0101] In one embodiment, the system further includes an algorithm parameter adjustment module, used to acquire historical image data of historical fire events in the target area and historical illumination data corresponding to the historical fire events; analyze the historical image data and historical illumination data to determine the correlation between the image judgment deviation of historical fire events and the illumination change pattern of the target area, so as to train the artificial intelligence model and obtain the correlation model; input the illumination information of the target area at the current moment into the correlation model to obtain the image judgment deviation corresponding to the illumination information at the current moment, and adjust the parameters of the image processing algorithm according to the image judgment deviation of the illumination information at the current moment.

[0102] In one embodiment, the fire extinguishing patch determination module 202 is further configured to acquire temperature distribution data of the target area, perform data preprocessing on the temperature distribution data to obtain a temperature distribution map of the target area; calculate the temperature gradient value at each location in the temperature distribution map, and determine the location where the temperature gradient value is greater than a preset gradient threshold as the ignition location; if the area of ​​the region formed by multiple consecutive ignition locations is greater than a preset area threshold, then it is determined that there is a fire point in the target area.

[0103] In one embodiment, the fire extinguishing patch determination module 202 is further configured to determine the location of the ignition point; determine at least one fire extinguishing patch from a plurality of fire extinguishing patches that is closest to the location of the ignition point as a working fire extinguishing patch; determine the location of the working fire extinguishing patch; and determine a first spray angle based on the location of the working fire extinguishing patch and the location of the ignition point.

[0104] In one embodiment, the spray angle adjustment module 204 is further configured to acquire a motion sequence image of the fire extinguishing particles sprayed at a first spray angle using a high-speed camera; determine the motion trajectory based on the motion sequence image; obtain the position of each frame corresponding to the motion trajectory in the motion sequence image; determine the expected motion trajectory of the fire extinguishing particles based on the position of the working fire extinguishing patch and the position of the ignition point; for each frame position, if the difference between the frame position and the corresponding point in the expected motion trajectory is greater than a preset deviation threshold, then determine the difference as the target difference value corresponding to the frame position; input all target differences into a pre-trained parameter adjustment model to obtain the parameter adjustment value output by the parameter adjustment model; and determine the second spray angle based on the parameter adjustment value and the first spray angle.

[0105] In one embodiment, the secondary fire suppression module is also used to acquire temperature change data and smoke concentration data at the ignition point; within a preset time period after the working fire suppression patch has finished spraying, if the current temperature value indicated by the temperature change data is less than a preset temperature threshold, and the current smoke concentration value indicated by the smoke concentration data is less than a preset concentration threshold, then it is determined that the fire at the ignition point has been extinguished.

[0106] In summary, the embodiments of this application have at least the following beneficial effects:

[0107] In this embodiment, firstly, based on the initial image data of the target area at the current moment, it is determined whether a potential fire source exists in the target area. If a potential fire source exists, then based on the temperature distribution data of the target area, it is determined whether an ignition point exists in the target area. When an ignition point exists, a working fire extinguishing patch and its first spray angle are determined based on the ignition point, wherein the working fire extinguishing patch is at least one of several fire extinguishing patches located in the target area. Then, the working fire extinguishing patch is driven to spray fire extinguishing particles toward the ignition point at the first spray angle. The trajectory of the fire extinguishing particles sprayed at the first spray angle is acquired, and a second spray angle is determined based on the trajectory, the position of the working fire extinguishing patch, and the position of the ignition point. The working fire extinguishing patch is then driven to spray fire extinguishing particles toward the ignition point at the second spray angle. This application uses data acquired by multiple sensors to accurately determine the occurrence of potential fires in the target area. Then, it determines the working fire extinguishing patch and its spray angle based on the ignition point. By analyzing the movement trajectory of the fire extinguishing particles, it adjusts the spray angle of the working fire extinguishing patch to control the fire extinguishing patch to accurately spray the inactivating particles towards the ignition point. This allows for precise control of the fire extinguishing patch to extinguish the burning material at the ignition point, effectively protecting important documents in the archives during a fire.

[0108] Through the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary hardware platforms, or it can be implemented entirely by hardware. Based on this understanding, all or part of the technical solutions of this application that contribute to the background technology can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of this application.

[0109] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A fire extinguishing control method based on fire extinguishing patches, characterized in that, The method includes: Based on the initial image data of the target area at the current moment, determine whether there is a potential fire source in the target area; If a potential fire source is found, the temperature distribution data of the target area is obtained to determine whether there is an ignition point in the target area. When there is an ignition point, the working fire extinguishing patch and its first spray angle are determined according to the ignition point. The working fire extinguishing patch is at least one of a number of fire extinguishing patches placed in the target area. Drive the working fire extinguishing patch to spray fire extinguishing particles toward the ignition point at the first spray angle; The motion trajectory of the fire extinguishing particles ejected at the first spray angle is obtained, and a second spray angle is determined based on the motion trajectory, the position of the working fire extinguishing patch and the position of the ignition point, and the working fire extinguishing patch is driven to spray fire extinguishing particles toward the ignition point at the second spray angle; The step of determining whether a potential fire source exists in the target area based on the initial image data of the target area at the current moment includes: acquiring the initial image data of the target area at the current moment; performing image preprocessing on the initial image data using an image processing algorithm to obtain image data to be tested; extracting image features from the image data to be tested to obtain feature vectors of the image data to be tested, and inputting the feature vectors into a pre-trained feature classification model to obtain multiple types of image features; if the multiple types of image features include fire light wave features, then a potential fire source is determined to exist; wherein, the fire light wave features are determined based on historical fire image data; The parameters of the image processing algorithm are adjusted through the following steps: Acquire historical image data of historical fire events in the target area and historical illumination data corresponding to the historical fire events; The historical image data and historical illumination data are analyzed to determine the correlation between the image judgment deviation of the historical fire event and the illumination change pattern of the target area, so as to train the artificial intelligence model and obtain the correlation model. The current illumination information of the target area is input into the association model to obtain the image judgment deviation corresponding to the current illumination information, and the parameters of the image processing algorithm are adjusted according to the image judgment deviation of the current illumination information.

2. The fire extinguishing control method based on fire extinguishing patches as described in claim 1, characterized in that, Also includes: Based on the obtained temperature change data and smoke concentration data at the ignition point, it is determined whether the fire at the ignition point has been extinguished; If the fire at the ignition point is not extinguished, and the current temperature value of the ignition point indicated by the temperature change data is greater than the preset temperature, then at least one fire extinguishing patch adjacent to the working fire extinguishing patch among the plurality of fire extinguishing patches is driven to spray fire extinguishing particles toward the ignition point, or the output end of the inert gas generator corresponding to the ignition point is driven to release inert gas toward the ignition point.

3. The fire extinguishing control method based on fire extinguishing patches as described in claim 2, characterized in that, After the steps of driving at least one fire extinguishing patch adjacent to the working fire extinguishing patch in the plurality of fire extinguishing patches to spray fire extinguishing particles toward the ignition point, or driving the output end of an inert gas generator corresponding to the ignition point to release inert gas toward the ignition point, the method further includes: The real-time temperature and smoke concentration at the ignition point are obtained to determine whether the fire at the ignition point has been extinguished. If the fire at the ignition point has been extinguished, the ventilation system corresponding to the target area is activated.

4. The fire extinguishing control method based on fire extinguishing patches as described in claim 1, characterized in that, The step of determining whether there is an ignition point in the target area based on the obtained temperature distribution data of the target area includes: Acquire temperature distribution data of the target area, and perform data preprocessing on the temperature distribution data to obtain a temperature distribution map of the target area; Calculate the temperature gradient value at each location in the temperature distribution map, and determine the location where the temperature gradient value is greater than a preset gradient threshold as the ignition location; If the area formed by multiple consecutive ignition points is greater than a preset area threshold, then it is determined that there is a fire point in the target area.

5. The fire extinguishing control method based on fire extinguishing patches as described in claim 1, characterized in that, The step of determining the working fire extinguishing patch and its first spray angle based on the ignition point includes: Determine the location of the ignition point; From the plurality of fire extinguishing stickers, at least one fire extinguishing sticker that is closest to the location of the fire point is identified as the working fire extinguishing sticker; The position of the working fire extinguishing patch is determined, and the first spray angle is determined based on the position of the working fire extinguishing patch and the position of the ignition point.

6. The fire extinguishing control method based on fire extinguishing patches as described in claim 1, characterized in that, The step of acquiring the trajectory of the extinguishing particles ejected at the first spray angle, and determining the second spray angle based on the trajectory, the position of the working extinguishing patch, and the position of the ignition point, includes: The motion sequence image of the fire extinguishing particles sprayed at the first spray angle is captured by a high-speed camera. The motion trajectory is determined based on the motion sequence image, and the position of each frame corresponding to the motion trajectory is obtained in the motion sequence image. The expected trajectory of the extinguishing particles is determined based on the position of the working fire extinguishing patch and the position of the ignition point. For each frame position, if the difference between the frame position and the corresponding point in the expected motion trajectory is greater than a preset deviation threshold, then the difference is determined to be the target difference corresponding to the frame position. All the target differences are input into the pre-trained parameter adjustment model to obtain the parameter adjustment values ​​output by the parameter adjustment model, and the second injection angle is determined based on the parameter adjustment values ​​and the first injection angle.

7. The fire extinguishing control method based on fire extinguishing patches as described in claim 2, characterized in that, The step of determining whether the fire at the ignition point has been extinguished based on the obtained temperature change data and smoke concentration data of the ignition point includes: Acquire temperature change data and smoke concentration data at the ignition point; If, within a preset time period after the working fire extinguishing patch has finished spraying, the current temperature value indicated by the temperature change data is less than a preset temperature threshold, and the current smoke concentration value indicated by the smoke concentration data is less than a preset concentration threshold, then it is determined that the fire at the ignition point has been extinguished.

8. A fire extinguishing control system based on fire extinguishing patches, characterized in that, include: The potential ignition source detection module is used to determine whether a potential ignition source exists in the target area based on the initial image data of the target area at the current moment. The fire extinguishing patch determination module is used to determine whether there is an ignition point in the target area based on the temperature distribution data of the target area if a potential fire source is found. When there is an ignition point, the working fire extinguishing patch and its first spray angle are determined based on the ignition point. The working fire extinguishing patch is at least one of a plurality of fire extinguishing patches disposed in the target area. A fire extinguishing patch driving module is used to drive the working fire extinguishing patch to spray fire extinguishing particles toward the ignition point at the first spray angle. The spray angle adjustment module is used to acquire the movement trajectory of the fire extinguishing particles sprayed at the first spray angle, and determine the second spray angle based on the movement trajectory, the position of the working fire extinguishing patch and the position of the ignition point, and drive the working fire extinguishing patch to spray fire extinguishing particles toward the ignition point at the second spray angle; The step of determining whether a potential fire source exists in the target area based on the initial image data of the target area at the current moment includes: acquiring the initial image data of the target area at the current moment; performing image preprocessing on the initial image data using an image processing algorithm to obtain image data to be tested; extracting image features from the image data to be tested to obtain feature vectors of the image data to be tested, and inputting the feature vectors into a pre-trained feature classification model to obtain multiple types of image features; if the multiple types of image features include fire light wave features, then a potential fire source is determined to exist; wherein, the fire light wave features are determined based on historical fire image data; The parameters of the image processing algorithm are adjusted through the following steps: Acquire historical image data of historical fire events in the target area and historical illumination data corresponding to the historical fire events; The historical image data and historical illumination data are analyzed to determine the correlation between the image judgment deviation of the historical fire event and the illumination change pattern of the target area, so as to train the artificial intelligence model and obtain the correlation model. The current illumination information of the target area is input into the association model to obtain the image judgment deviation corresponding to the current illumination information, and the parameters of the image processing algorithm are adjusted according to the image judgment deviation of the current illumination information.

Images