Hospital intelligent fire-fighting monitoring system and method based on Internet of Things

Abstract

The invention discloses a hospital intelligent fire-fighting monitoring system and method based on the Internet of Things, and relates to the technical field of the Internet of Things. Spatial position coordinates of fire extinguishing equipment, a sensor and a monitoring camera are acquired; training a fire behavior prediction model; predicting a probability value that the internal image of the hospital belongs to each fire category by using a fire prediction model, judging whether a fire exists, and if so, triggering first-level early warning; acquiring environmental parameters in continuous n1 normal sampling intervals to calculate a comprehensive fire risk value, and if the comprehensive fire risk value is less than or equal to a risk threshold, determining that no fire exists; if the comprehensive fire risk value is greater than a risk threshold value, determining that a fire exists, and triggering second-level early warning; calculating an adjustment coefficient, obtaining a second sampling interval according to the adjustment coefficient, and sampling abnormal environment parameters; the normal environment parameters are obtained, the difference value is calculated, if the difference value is larger than the difference value threshold value, the spatial position coordinates of the sensor serve as the fire point, the fire extinguishing equipment is started, and the fire behavior recognition accuracy is improved.

Description

Technical Field

[0001] The present application relates to the field of Internet of Things technology, and in particular to a hospital smart fire monitoring system and method based on the Internet of Things. Background Art

[0002] Hospitals are places with dense personnel, complex equipment and high fire risks. Traditional fire monitoring mainly relies on manual inspections and smoke alarms, which have problems such as untimely response and many blind spots. The development of Internet of Things technology has brought new solutions to hospital fire safety.

[0003] Existing fire monitoring mainly uses the following types of IoT technologies: wireless sensor networks, video monitoring and image recognition, big data analysis and early warning, and IoT integrated management platforms. Various fire detectors are used to monitor the internal environment of buildings in real time. Once a fire occurs, it can be detected and triggered in time to notify people to evacuate, thus buying precious time for rescue.

[0004] However, existing fire monitoring systems often only use a single early warning mechanism, which has the risk of failure and is prone to missed or false alarms. Summary of the invention

[0005] The present application aims to solve at least one of the technical problems in the related art to a certain extent. To this end, one purpose of the present application is to propose a hospital smart fire monitoring system and method based on the Internet of Things to improve the accuracy of fire identification.

[0006] One aspect of the present application provides a hospital smart fire monitoring method based on the Internet of Things, including:

[0007] Step S100: Acquire the spatial position coordinates of fire extinguishing equipment, sensors and surveillance cameras;

[0008] The specific method for obtaining the spatial position coordinates of the fire extinguishing equipment, sensors and surveillance cameras is:

[0009] Step S110: deploying fire extinguishing equipment, sensor networks and surveillance camera networks in key areas within the hospital, wherein the sensor networks include temperature sensors, smoke sensors and toxic gas sensors;

[0010] Step S120: setting a sampling interval mechanism for sensors and surveillance cameras, the sampling interval mechanism is switched according to the triggering and release of the first level warning and the second level warning, the sampling interval mechanism includes a normal sampling interval and a second sampling interval, the sensor network and the surveillance camera network use the normal sampling interval as the sampling interval mechanism in the absence of fire, when there is no fire and the first level warning is triggered, the normal sampling interval is used as the sampling interval mechanism, and when the second level warning is triggered, the sampling interval mechanism is switched to the second sampling interval;

[0011] Step S130: construct a three-dimensional model of the interior of the hospital, use a laser radar to perform a comprehensive scan of the interior of the hospital, obtain point cloud data of fire extinguishing equipment, sensor point cloud data and camera point cloud data, and use the center point position of each fire extinguishing equipment point cloud, sensor point cloud and camera point cloud as the spatial position coordinates of the corresponding fire extinguishing equipment, sensor and surveillance camera;

[0012] Step S200: obtaining N historical hospital interior images, marking the fire category of each historical hospital interior image, wherein the fire category includes fire and no fire, and using the marked historical hospital interior images to train a fire prediction model;

[0013] The method of obtaining N historical hospital interior images, marking the fire category of each historical hospital interior image, wherein the fire category includes fire and no fire, and using the marked historical hospital interior images to train the fire prediction model is as follows:

[0014] Step S210: Obtain N historical hospital interior images from the hospital's monitoring fire database, wherein the historical hospital interior images include hospital interior images of different areas in the hospital and different historical time periods;

[0015] Step S220: define a fire category label, 0 means no fire, 1 means fire, and label each historical hospital interior image with a fire category. If there is a fire feature, it is labeled as 1, otherwise it is labeled as 0;

[0016] Step S230: using the annotated historical hospital internal images as training samples, constructing a structure of a convolutional neural network model, wherein the convolutional neural network model includes two convolutional layers, two maximum pooling layers, and two fully connected layers;

[0017] Step S240: using the cross entropy loss function as the loss function and the Adam optimizer as the optimizer, in each training round, using the historical hospital interior images as input data, using the convolutional neural network model to predict the fire category of the historical hospital interior images, outputting the corresponding fire category label, calculating the loss function value, and updating the model parameters according to the back propagation algorithm;

[0018] Step S250: Taking minimizing the loss function value between the predicted fire category label and the actual fire category label as the training goal, when the loss function value reaches convergence, the training is completed and a trained fire prediction model is obtained.

[0019] Step S300: Obtain an image of the interior of the hospital, use the fire prediction model to predict the probability value of the image belonging to each fire category, and determine whether there is a fire based on the probability value. If there is a fire, trigger the first level warning;

[0020] The specific method of obtaining the hospital interior image, using the fire prediction model to predict the probability value of the hospital interior image belonging to each fire category, judging whether there is a fire according to the probability value, and triggering the first level warning if there is a fire is as follows:

[0021] Step S310: setting a normal sampling interval f0, where the normal sampling interval is the sampling interval of the surveillance camera and the sensor when there is no fire. During each normal sampling interval, the surveillance camera inside the hospital performs real-time sampling to obtain an image of the inside of the hospital.

[0022] Step S320: using the hospital interior image as input data, using the fire prediction model to predict the probability value of the hospital interior image belonging to the two fire categories of fire and no fire. If the probability value of the output fire category being fire is less than p1%, continue to maintain the normal sampling interval to sample the surveillance camera and sensor;

[0023] Step S330: If the output fire category has a probability value of fire greater than or equal to p1%, a first level warning is triggered.

[0024] Step S400: within n1 consecutive normal sampling intervals after the first level warning is triggered, the environmental parameters are obtained through sensors and the comprehensive fire risk value is calculated. If the comprehensive fire risk value of the n1 normal sampling intervals is less than or equal to the risk threshold, it is confirmed that there is no fire and the first level warning is lifted;

[0025] In the continuous n1 normal sampling intervals after the first level warning is triggered, the environmental parameters are obtained through sensors and the comprehensive fire risk value is calculated. If the comprehensive fire risk value of the n1 normal sampling intervals is less than or equal to the risk threshold, it is confirmed that there is no fire, and the specific method of lifting the first level warning is:

[0026] Step S410: for the hospital interior image predicted to have a fire, find the surveillance camera that acquired the hospital interior image, and use the spatial position coordinates of the surveillance camera as the fire shooting point;

[0027] Step S420: Calculate the distance between the fire shooting point (x, y, z) and the spatial position coordinates of each sensor, and find the temperature sensor, smoke sensor and poison gas sensor closest to the fire shooting point;

[0028] Step S430: Set the number of continuous sampling times n1. Within n1 normal sampling intervals after the first-level warning is triggered, the temperature sensor, smoke sensor and toxic gas sensor are used to sample and obtain environmental parameters, wherein the environmental parameters include ambient temperature, smoke concentration and toxic gas concentration. The local temperature distribution T = {T1, T2, ..., T n1}、Local smoke concentration distribution S={S1,S2,...,Sn1}、Local gas concentration distribution G={G1,G2,...,G n1}, where T n1 , S n1 , G n1 They represent the ambient temperature, smoke concentration, and toxic gas concentration of the n1th normal sampling interval respectively;

[0029] Step S440: within n1 normal sampling intervals after the first level warning is triggered, calculating the comprehensive fire risk value of the environmental parameters sampled in the normal sampling intervals;

[0030] The calculation method of the comprehensive fire risk value is:

[0031] Step S441: Calculate the average ambient temperature T in n1 normal sampling intervals according to the local temperature distribution, local smoke concentration distribution and local toxic gas concentration distribution. avg , average smoke density S avg and the average toxic gas concentration G avg ;

[0032] Step S442: Acquire normal environmental parameters, which include normal environmental temperature T0, normal smoke concentration S0 and normal toxic gas concentration G0, combined with the average environmental temperature T avg , average smoke density S avg and the average toxic gas concentration G avg Calculate the temperature risk value R of the i-th normal sampling interval T,i , smoke risk value R S,i and the toxic gas risk value R G,i ;

[0033] Step S443: Calculate the fire risk value H of the ith normal sampling interval based on the temperature risk value, smoke risk value and toxic gas risk value of the ith normal sampling interval. i ;

[0034] Step S444: Calculate the comprehensive fire risk value H of n1 normal sampling intervals according to the fire risk value of each normal sampling interval in n1 normal sampling intervals. total ;

[0035] Step S450: Preset risk threshold H th , compare the comprehensive fire risk value within n1 normal sampling intervals with the risk threshold. If the comprehensive fire risk value within n1 normal sampling intervals is less than or equal to the risk threshold, it is confirmed that there is no fire and the first-level warning is lifted;

[0036] Step S500: If the comprehensive fire risk value of n1 normal sampling intervals is greater than the risk threshold, it is confirmed that there is a fire and the second-level warning is triggered; the adjustment coefficient is calculated according to the comprehensive fire risk value, and the second sampling interval is obtained according to the adjustment coefficient. The sensor samples the abnormal environmental parameters according to the second sampling interval;

[0037] The specific method of calculating the adjustment coefficient according to the comprehensive fire risk value, obtaining the second sampling interval according to the adjustment coefficient, and the sensor sampling the abnormal environment parameter according to the second sampling interval is:

[0038] Step S510: Set the range of the adjustment coefficient k to obtain the minimum adjustment coefficient k min and the maximum adjustment factor k max ;

[0039] Step S520: Calculate the difference k between the maximum adjustment coefficient and the minimum adjustment coefficient max -k min , calculate the risk excess of the comprehensive fire risk value compared to the risk threshold H total -H th , divide the excess risk by the risk threshold to get the proportionality coefficient The proportional coefficient is multiplied by the difference to obtain the adjustment amount, and the adjustment amount is added to the minimum adjustment coefficient to obtain the adjustment coefficient k;

[0040] Step S530: multiplying the adjustment coefficient k by the normal sampling interval f0 to obtain a second sampling interval f2;

[0041] Step S540: The sensor samples abnormal environmental parameters according to the second sampling interval, wherein the abnormal environmental parameters include abnormal environmental temperature T yc , Abnormal smoke density S yc , Abnormal toxic gas concentration G yc .

[0042] Step S600: obtaining normal environmental parameters, and calculating a difference value in combination with abnormal environmental parameters. If the difference value is greater than a difference threshold value within n2 consecutive second sampling intervals, the spatial position coordinates of the sensor that obtains the abnormal environmental parameters are used as the fire point, and the N fire extinguishing devices closest to the fire point are activated;

[0043] The specific method of obtaining the normal environmental parameters and calculating the difference value in combination with the abnormal environmental parameters, and if the difference value is greater than the difference threshold value within n2 consecutive second sampling intervals, taking the spatial position coordinates of the sensor that obtains the abnormal environmental parameters as the fire point, and activating the N fire extinguishing devices closest to the fire point is:

[0044] Step S610: acquiring normal environmental parameters, the sensor acquires abnormal environmental parameters within n2 consecutive second sampling intervals after the second-level warning is triggered, and calculating the difference between the abnormal environmental parameters acquired by the sensor and the normal environmental parameters, the normal environmental parameters including: normal environmental temperature T0, normal smoke concentration S0 and normal toxic gas concentration G0, and the difference between the abnormal environmental parameters acquired by the sensor and the normal environmental parameters includes the difference value of environmental temperature, the difference value of smoke concentration and the difference value of toxic gas concentration;

[0045] Step S620: preset difference thresholds, the difference thresholds including temperature difference thresholds, smoke difference thresholds and poison gas difference thresholds, compare the temperature difference threshold with the ambient temperature difference value, the smoke difference threshold with the smoke concentration difference value, and the poison gas difference threshold with the poison gas concentration difference value, if the difference value is continuously greater than the difference threshold value within n2 consecutive second sampling intervals, then take the spatial position coordinates of the sensor that obtains the abnormal environmental parameters as the fire point;

[0046] Step S630: Calculate the distance between the fire point and all fire-fighting equipment inside the hospital, and activate the N fire-fighting equipment closest to it.

[0047] One aspect of the present application provides a hospital smart fire monitoring system based on the Internet of Things, including:

[0048] IoT device deployment module, used to obtain the spatial location coordinates of fire-fighting equipment, sensors, and surveillance cameras;

[0049] A prediction model training module is used to obtain N historical hospital interior images, annotate the fire category of each historical hospital interior image, wherein the fire category includes fire and no fire, and train a fire prediction model using the annotated historical hospital interior images;

[0050] The first-level warning trigger module is used to obtain images of the interior of the hospital, use the fire prediction model to predict the probability value of the hospital interior image belonging to each fire category, and determine whether there is a fire based on the probability value. If there is a fire, the first-level warning is triggered;

[0051] The first-level fire confirmation module is used to obtain environmental parameters and calculate the comprehensive fire risk value through sensors within n1 consecutive normal sampling intervals after the first-level warning is triggered. If the comprehensive fire risk value of n1 normal sampling intervals is less than or equal to the risk threshold, it is confirmed that there is no fire and the first-level warning is lifted;

[0052] The second-level fire confirmation module is used to confirm that there is a fire and trigger the second-level warning if the comprehensive fire risk value of n1 normal sampling intervals is greater than the risk threshold; calculate the adjustment coefficient according to the comprehensive fire risk value, obtain the second sampling interval according to the adjustment coefficient, and the sensor samples the abnormal environmental parameters according to the second sampling interval;

[0053] The fire extinguishing equipment startup module is used to obtain normal environmental parameters and calculate the difference value in combination with abnormal environmental parameters. If the difference value is greater than the difference threshold within n2 consecutive second sampling intervals, the spatial position coordinates of the sensor that obtains the abnormal environmental parameters are used as the fire point, and the N fire extinguishing equipment closest to the fire point are started.

[0054] One aspect of the present application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps in the hospital smart fire monitoring method based on the Internet of Things are implemented.

[0055] One aspect of the present application provides a readable storage medium, which stores a computer program, and the computer program is suitable for loading by a processor to execute the steps in a hospital smart fire monitoring method based on the Internet of Things.

[0056] Compared with the existing technology, the hospital smart fire monitoring system and method based on the Internet of Things proposed in this application have the following advantages:

[0057] This application adopts a method that combines a visual-based fire prediction model with an early warning mechanism based on sensor-based environmental parameter analysis. The visual model makes a preliminary judgment on the fire situation, which is further verified by sensor data. It effectively overcomes the limitations of a single early warning, reduces the risk of missed reports and false alarms, and improves the reliability of the early warning.

[0058] This application introduces a two-level warning mechanism. When the fire prediction model finds a suspected fire, it triggers the first level warning, and then quickly starts the sensor data collection and analysis. There is no need to wait until the sensor detects obvious environmental anomalies before issuing a warning, but the warning time point is moved forward. At the same time, the sampling interval is also dynamically adjusted according to the risk level. When the risk is high, the sampling is more frequent, and the overall warning response is faster.

[0059] This application perceives fire risks from two dimensions: images and environmental parameters, covering different senses of vision and smell. Different from a single perception path, this solution comprehensively perceives fire hazards everywhere inside the building, leaving no blind spots.

[0060] The sensor layout is integrated with the three-dimensional model, and the spatial coordinates of each sensor are known. When the abnormal environmental parameters continue to exceed the standard, the coordinates of the corresponding sensor can be determined as the fire point, and the fire location can be quickly locked. Combined with the preset fire extinguishing equipment location information, precise and targeted rescue can be carried out to improve fire extinguishing efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 A method flow chart of the hospital smart fire monitoring method based on the Internet of Things provided in this application;

[0062] Figure 2 A two-level warning mechanism flow chart of the hospital smart fire monitoring method based on the Internet of Things provided in this application;

[0063] Figure 3 Functional module diagram of the hospital smart fire monitoring system based on the Internet of Things provided for this application;

[0064] Figure 4 A schematic diagram of the structure of an electronic device provided in this application;

[0065] Figure 5 It is a schematic diagram of the structure of a readable storage medium provided by this application. DETAILED DESCRIPTION

[0066] In order to better understand the present application, various aspects of the present application will be described in more detail with reference to the accompanying drawings. It should be understood that these detailed descriptions are only descriptions of exemplary embodiments of the present application and do not limit the scope of the present application in any way. Throughout the specification, the same reference numerals refer to the same elements. The expression "and / or" includes any and all combinations of one or more of the associated listed items.

[0067] In the accompanying drawings, the size, dimensions and shapes of the elements have been slightly adjusted for ease of illustration. The drawings are for illustration only and are not strictly drawn to scale. As used herein, the terms "substantially", "approximately" and similar terms are used as terms of approximation, not as terms of degree, and are intended to illustrate the inherent deviations in measurements or calculations that will be recognized by those of ordinary skill in the art. In addition, in this application, the order in which the steps are described does not necessarily represent the order in which these processes occur in actual operation, unless otherwise specified or can be derived from the context.

[0068] It should also be understood that expressions such as "including", "comprising", "having", "containing" and / or "comprising" are open rather than closed expressions in this specification, which indicate the presence of the stated features, elements and / or components, but do not exclude the presence of one or more other features, elements, components and / or combinations thereof. In addition, when expressions such as "at least one of..." appear after a list of listed features, they modify the entire list of features rather than just the individual elements in the list. In addition, when describing embodiments of the present application, "may" is used to mean "one or more embodiments of the present application". And, the term "exemplary" is intended to refer to an example or illustration.

[0069] Unless otherwise specified, all words (including engineering terms and scientific and technological terms) used in this article have the same meaning as those commonly understood by ordinary technicians in the field to which this application belongs. It should also be understood that, unless clearly stated in this application, words defined in common dictionaries should be interpreted as having the same meaning as their meaning in the context of the relevant technology, and should not be interpreted in an idealized or overly formal sense.

[0070] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0071] Example 1

[0072] like Figure 1 As shown, the hospital smart fire monitoring method based on the Internet of Things provided by this application includes:

[0073] Step S100: Acquire the spatial position coordinates of fire extinguishing equipment, sensors and surveillance cameras;

[0074] The specific method for obtaining the spatial position coordinates of the fire extinguishing equipment, sensors and surveillance cameras is:

[0075] Step S110: deploying fire extinguishing equipment, sensor networks and surveillance camera networks in key areas within the hospital, wherein the sensor networks include temperature sensors, smoke sensors and toxic gas sensors;

[0076] Step S120: setting a sampling interval mechanism for sensors and surveillance cameras, the sampling interval mechanism is switched according to the triggering and release of the first level warning and the second level warning, the sampling interval mechanism includes a normal sampling interval and a second sampling interval, the sensor network and the surveillance camera network use the normal sampling interval as the sampling interval mechanism in the absence of fire, when there is no fire and the first level warning is triggered, the normal sampling interval is used as the sampling interval mechanism, and when the second level warning is triggered, the sampling interval mechanism is switched to the second sampling interval;

[0077] Step S130: Use laser radar to perform a comprehensive scan of the interior of the hospital, build a three-dimensional model of the interior of the hospital, obtain point cloud data of fire extinguishing equipment, sensor point cloud data, and camera point cloud data, and use the center point position of each fire extinguishing equipment point cloud, sensor point cloud, and camera point cloud as the spatial position coordinates of the corresponding fire extinguishing equipment, sensor, and surveillance camera;

[0078] The sensor layout is integrated with the three-dimensional model, and the spatial position coordinates of each sensor are known. When the environmental parameters continue to be abnormal, the spatial position coordinates of the corresponding sensor can be determined as the fire point, and the fire location can be quickly locked. Combined with the spatial position coordinates of the preset fire-fighting equipment, it is convenient to carry out precise and targeted rescue in the future, improving the efficiency of fire-fighting.

[0079] Step S200: obtaining N historical hospital interior images, marking the fire category of each historical hospital interior image, wherein the fire category includes fire and no fire, and using the marked historical hospital interior images to train a fire prediction model;

[0080] The method of obtaining N historical hospital interior images, marking the fire category of each historical hospital interior image, wherein the fire category includes fire and no fire, and using the marked historical hospital interior images to train the fire prediction model is as follows:

[0081] Step S210: Obtain N historical hospital interior images from the hospital's monitoring fire database, wherein the historical hospital interior images include hospital interior images of different areas in the hospital and different historical time periods;

[0082] The N historical hospital interior images include historical hospital interior images with fire and historical hospital interior images without fire, wherein the historical hospital interior images with fire include historical actual fire images and software simulated fire images;

[0083] The method of acquiring the historical actual fire image is: by retrieving the historical fire accident monitoring images recorded by the hospital fire protection system, extracting the key frame images when the fire occurred;

[0084] The software simulated fire image is obtained by: using 3D modeling software, building a virtual hospital scene model according to the internal structure layout of the hospital, simulating the fire process in the virtual scene, controlling the fire source location, flame size, and smoke concentration, and rendering to generate the software simulated fire image;

[0085] Step S220: define a fire category label, 0 means no fire, 1 means fire, and label each historical hospital interior image with a fire category. If there is a fire feature, it is labeled as 1, otherwise it is labeled as 0;

[0086] The fire characteristics include: open flames, thick smoke;

[0087] Step S230: using the annotated historical hospital internal images as training samples, constructing a structure of a convolutional neural network model, wherein the convolutional neural network model includes two convolutional layers, two maximum pooling layers, and two fully connected layers;

[0088] The convolution layer is used to extract local features of historical hospital internal images to obtain feature maps, the pooling layer is used to reduce the resolution of feature maps, and the fully connected layer is used to perform feature fusion and classification prediction on feature maps;

[0089] Step S240: using the cross entropy loss function as the loss function and the Adam optimizer as the optimizer, in each training round, using the historical hospital interior images as input data, using the convolutional neural network model to predict the fire category of the historical hospital interior images, outputting the corresponding fire category label, calculating the loss function value, and updating the model parameters according to the back propagation algorithm;

[0090] Step S250: Taking the loss function value between the predicted fire category label and the actual fire category label as the training target, when the loss function value reaches convergence, the training is completed and a trained fire prediction model is obtained;

[0091] The calculation formula of the loss function is: Among them, y n represents the true fire category label of the historical hospital interior image, represents the predicted probability of belonging to the nth fire category label, log(·) is the logarithmic function, N lb Indicates the number of categories of fire category labels, N lb is equal to 2;

[0092] y n =1 means that the training sample belongs to the nth fire category label, y n =0 means that the training sample does not belong to the nth fire category label;

[0093] Step S300: Obtain an image of the interior of the hospital, use the fire prediction model to predict the probability value of the image belonging to each fire category, and determine whether there is a fire based on the probability value. If there is a fire, trigger the first level warning;

[0094] The specific method of obtaining the hospital interior image, using the fire prediction model to predict the probability value of the hospital interior image belonging to each fire category, judging whether there is a fire according to the probability value, and triggering the first level warning if there is a fire is as follows:

[0095] Step S310: setting a normal sampling interval f0, where the normal sampling interval is the sampling interval of the surveillance camera and the sensor when there is no fire. During each normal sampling interval, the surveillance camera inside the hospital performs real-time sampling to obtain an image of the inside of the hospital.

[0096] The normal sampling interval is set by those skilled in the art according to actual needs;

[0097] The collection area of ​​the hospital internal images should cover the key areas inside the hospital, including wards, corridors, and warehouses;

[0098] Step S320: using the hospital interior image as input data, using the fire prediction model to predict the probability value of the hospital interior image belonging to the two fire categories of fire and no fire. If the probability value of the output fire category being fire is less than p1%, continue to maintain the normal sampling interval to sample the surveillance camera and sensor;

[0099] Step S330: If the output fire category has a probability value of fire greater than or equal to p1%, a first-level warning is triggered;

[0100] The p1% is a first warning threshold, and the value of the first warning threshold is set by those skilled in the art according to actual needs;

[0101] If the probability value of fire is greater than the first warning threshold, a first-level warning is triggered, and the warning time and the location of the surveillance camera that obtains the internal image of the hospital are recorded;

[0102] If the probability value of fire is less than the first warning threshold, it is not considered that there is a fire in the hospital internal image, and the first level warning is not triggered;

[0103] The vision-based fire prediction model is used to make a preliminary judgment on the fire situation. However, in order to avoid misjudgment and missed judgment caused by a single judgment standard, the prediction results of the fire prediction model are used as the basis for the first-level warning, and then the sensor data collection and analysis are quickly started. The data analysis structure collected by the sensor is used as the basis for further verification of the fire situation. The dual mechanisms complement each other.

[0104] Step S400: within n1 consecutive normal sampling intervals after the first level warning is triggered, the environmental parameters are obtained through sensors and the comprehensive fire risk value is calculated. If the comprehensive fire risk value of the n1 normal sampling intervals is less than or equal to the risk threshold, it is confirmed that there is no fire and the first level warning is lifted;

[0105] In the continuous n1 normal sampling intervals after the first level warning is triggered, the environmental parameters are obtained through sensors and the comprehensive fire risk value is calculated. If the comprehensive fire risk value of the n1 normal sampling intervals is less than or equal to the risk threshold, it is confirmed that there is no fire, and the specific method of lifting the first level warning is:

[0106] Step S410: for the hospital interior image predicted to have a fire, find the surveillance camera that acquired the hospital interior image, and use the spatial position coordinates of the surveillance camera as the fire shooting point;

[0107] Step S420: Calculate the distance between the fire shooting point (x, y, z) and the spatial position coordinates of each sensor, and find the temperature sensor, smoke sensor and poison gas sensor closest to the fire shooting point;

[0108] Step S430: Set the number of continuous sampling times n1. Within n1 normal sampling intervals after the first-level warning is triggered, the temperature sensor, smoke sensor and toxic gas sensor are used to sample and obtain environmental parameters, wherein the environmental parameters include ambient temperature, smoke concentration and toxic gas concentration. The local temperature distribution T = {T1, T2, ..., T n1}、Local smoke concentration distribution S={S1,S2,...,S n1}、Local gas concentration distribution G={G1,G2,...,G n1}, where T n1 , S n1 , G n1 They represent the ambient temperature, smoke concentration, and toxic gas concentration of the n1th normal sampling interval respectively;

[0109] Step S440: within n1 normal sampling intervals after the first level warning is triggered, calculating the comprehensive fire risk value of the environmental parameters sampled in the normal sampling intervals;

[0110] The calculation method of the comprehensive fire risk value is:

[0111] Step S441: Calculate the average ambient temperature T in n1 normal sampling intervals according to the local temperature distribution, local smoke concentration distribution and local toxic gas concentration distribution. avg , average smoke density S avg and the average toxic gas concentration G avg ;

[0112] The calculation formula for the average ambient temperature within the n1 normal sampling intervals is: Among them, T i is the ambient temperature of the ith normal sampling interval;

[0113] The calculation formula for the average smoke concentration within the n1 normal sampling intervals is: Among them, S i is the smoke concentration of the ith normal sampling interval;

[0114] The calculation formula for the average toxic gas concentration within the n1 normal sampling intervals is: Among them, G i is the toxic gas concentration in the ith normal sampling interval;

[0115] Step S442: Acquire normal environmental parameters, which include normal environmental temperature T0, normal smoke concentration S0 and normal toxic gas concentration G0, combined with the average environmental temperature T avg , average smoke density S avg and the average toxic gas concentration G avg Calculate the temperature risk value R of the i-th normal sampling interval T,i , smoke risk value R S,i and the toxic gas risk value R G,i ;

[0116] The calculation formula of the temperature risk value is: Among them, a T represents the contribution weight of the deviation of ambient temperature to the temperature risk, k T is the contribution weight of the ambient temperature change rate to the temperature risk, ΔT i,i-1 represents the ambient temperature difference between the ith normal sampling interval and the i-1th normal sampling interval, Δt represents the sampling time difference between the ith normal sampling interval and the i-1th normal sampling interval;

[0117] The contribution weight of the ambient temperature deviation to the temperature risk and the contribution weight of the ambient temperature change rate to the temperature risk are set by those skilled in the art according to actual needs and experience;

[0118] The calculation formula of the smoke risk value is: Among them, a S represents the contribution weight of the deviation of smoke concentration to the smoke risk, k S is the contribution weight of smoke concentration change rate to smoke risk, ΔS i,i-1 It represents the smoke concentration difference between the i-th normal sampling interval and the i-1-th normal sampling interval;

[0119] The contribution weight of the smoke concentration deviation to the smoke risk and the contribution weight of the smoke concentration change rate to the smoke risk are set by those skilled in the art according to actual needs and experience;

[0120] The calculation formula of the toxic gas risk value is: Among them, a G represents the contribution weight of the deviation of toxic gas concentration to the toxic gas risk, k G is the contribution weight of the change rate of toxic gas concentration to the toxic gas risk, ΔG i,i-1 It represents the difference in toxic gas concentration between the i-th normal sampling interval and the i-1-th normal sampling interval;

[0121] The contribution weight of the deviation of the toxic gas concentration to the toxic gas risk and the contribution weight of the toxic gas concentration change rate to the toxic gas risk are set by those skilled in the art according to actual needs and experience;

[0122] The normal ambient temperature is determined based on real-time meteorological data;

[0123] The normal smoke concentration and normal toxic gas concentration are determined according to indoor air quality standards;

[0124] Step S443: Calculate the fire risk value H of the ith normal sampling interval based on the temperature risk value, smoke risk value and toxic gas risk value of the ith normal sampling interval. i ;

[0125] The calculation formula for the fire risk value of the i-th normal sampling interval is:

[0126] Among them, R T,i , R S,i , R G,i are the ambient temperature, smoke concentration, and toxic gas concentration of the ith normal sampling interval, respectively; α, β, and γ are the weight coefficients of the temperature-smoke risk interaction term, the smoke-toxic gas risk interaction term, and the toxic gas-temperature risk interaction term, respectively;

[0127] The weight coefficients of the temperature-smoke risk interaction term, the smoke and toxic gas risk interaction term, and the toxic gas-temperature risk interaction term reflect the degree of influence of the interaction between different risk factors on the fire risk value;

[0128] Step S444: Calculate the comprehensive fire risk value H of n1 normal sampling intervals according to the fire risk value of each normal sampling interval in n1 normal sampling intervals. total ;

[0129] The calculation formula of the comprehensive fire risk value is: Where exp(·) is the exponential function and σ is the time decay factor;

[0130] The time decay factor is set by those skilled in the art according to actual needs and is used to control the influence of the historical fire risk value;

[0131] is a normal distribution function, used to calculate the weight of the fire risk value of the ith normal sampling interval in the comprehensive fire risk value, wherein the weight decays as the time distance between the ith normal sampling interval and the n1th normal sampling interval increases;

[0132] The time decay factor determines the rate at which the weight decays. The larger the time decay factor, the slower the weight decays, and the greater the impact of the fire risk value in historical time.

[0133] Step S450: Preset risk threshold H th , compare the comprehensive fire risk value within n1 normal sampling intervals with the risk threshold. If the comprehensive fire risk value within n1 normal sampling intervals is less than or equal to the risk threshold, it is confirmed that there is no fire and the first-level warning is lifted;

[0134] The risk threshold is set by those skilled in the art according to actual needs;

[0135] The first-level warning is based on a visual fire prediction model, which analyzes surveillance images to make a preliminary judgment on suspected fires. Once a suspicious situation is found, an early warning is triggered immediately. This method can detect hidden dangers in the early stages of a fire and gain valuable time for subsequent disposal.

[0136] After the first-level warning is triggered, the sensor is immediately activated to start data collection in preparation for subsequent accurate judgment of the fire situation. This "warning-collection" linkage mechanism ensures that fire-related data is obtained at the first time, improving the efficiency of warning and response.

[0137] Step S500: If the comprehensive fire risk value of n1 normal sampling intervals is greater than the risk threshold, it is confirmed that there is a fire and the second-level warning is triggered; the adjustment coefficient is calculated according to the comprehensive fire risk value, and the second sampling interval is obtained according to the adjustment coefficient. The sensor samples the abnormal environmental parameters according to the second sampling interval;

[0138] The specific method of calculating the adjustment coefficient according to the comprehensive fire risk value, obtaining the second sampling interval according to the adjustment coefficient, and the sensor sampling the abnormal environment parameter according to the second sampling interval is:

[0139] Step S510: Set the range of the adjustment coefficient k to obtain the minimum adjustment coefficient k min and the maximum adjustment factor k max ;

[0140] Step S520: Calculate the difference k between the maximum adjustment coefficient and the minimum adjustment coefficient max -k min , calculate the risk excess of the comprehensive fire risk value compared to the risk threshold H total -H th , divide the excess risk by the risk threshold to get the proportionality coefficient The proportional coefficient is multiplied by the difference to obtain the adjustment amount, and the adjustment amount is added to the minimum adjustment coefficient to obtain the adjustment coefficient k;

[0141] The calculation formula of the adjustment coefficient k is:

[0142] Step S530: multiplying the adjustment coefficient k by the normal sampling interval f0 to obtain a second sampling interval f2;

[0143] The calculation formula of the second sampling interval f2 is: f2=k×f0;

[0144] Step S540: The sensor samples abnormal environmental parameters according to the second sampling interval, wherein the abnormal environmental parameters include abnormal environmental temperature T yc , Abnormal smoke density S yc , Abnormal toxic gas concentration G yc ;

[0145] The second-level warning is based on the first-level warning. It calculates the fire risk value by analyzing the temperature, smoke, and toxic gas related environmental parameters collected by the environmental sensor. The second-level warning will be triggered only when the risk value continues to exceed the set threshold. This mechanism reduces the possibility of false alarms and improves the accuracy of warnings through quantitative analysis.

[0146] The second-level warning will intelligently activate nearby fire-fighting equipment based on the location of the fire point. Before the firefighters arrive, the nearby fire-fighting resources can be used to control the spread of the fire. This mechanism is equivalent to providing an intelligent line of defense in addition to manual rescue, improving rescue efficiency.

[0147] Figure 2 A two-level warning mechanism flow chart of the hospital smart fire monitoring method based on the Internet of Things provided in this application;

[0148] Step S600: obtaining normal environmental parameters, and calculating a difference value in combination with abnormal environmental parameters. If the difference value is greater than a difference threshold value within n2 consecutive second sampling intervals, the spatial position coordinates of the sensor that obtains the abnormal environmental parameters are used as the fire point, and the N fire extinguishing devices closest to the fire point are activated;

[0149] The specific method of obtaining the normal environmental parameters and calculating the difference value in combination with the abnormal environmental parameters, and if the difference value is greater than the difference threshold value within n2 consecutive second sampling intervals, taking the spatial position coordinates of the sensor that obtains the abnormal environmental parameters as the fire point, and activating the N fire extinguishing devices closest to the fire point is:

[0150] Step S610: acquiring normal environmental parameters, the sensor acquires abnormal environmental parameters within n2 consecutive second sampling intervals after the second-level warning is triggered, and calculating the difference between the abnormal environmental parameters acquired by the sensor and the normal environmental parameters, the normal environmental parameters including: normal environmental temperature T0, normal smoke concentration S0 and normal toxic gas concentration G0, and the difference between the abnormal environmental parameters acquired by the sensor and the normal environmental parameters includes the difference value of environmental temperature, the difference value of smoke concentration and the difference value of toxic gas concentration;

[0151] The calculation formula of the ambient temperature difference is: ΔT cy =T yc -T0;

[0152] The calculation formula of the smoke concentration difference value is: ΔS cy =S yc -S0;

[0153] The calculation formula of the difference value of the toxic gas concentration is: ΔG cy =G yc -G0;

[0154] Step S620: preset difference thresholds, the difference thresholds including temperature difference thresholds, smoke difference thresholds and poison gas difference thresholds, compare the temperature difference threshold with the ambient temperature difference value, the smoke difference threshold with the smoke concentration difference value, and the poison gas difference threshold with the poison gas concentration difference value, if the difference value is continuously greater than the difference threshold value within n2 consecutive second sampling intervals, then take the spatial position coordinates of the sensor that obtains the abnormal environmental parameters as the fire point;

[0155] Step S630: Calculate the distance between the fire point and all fire-fighting equipment inside the hospital, and activate the N fire-fighting equipment closest to it.

[0156] Example 2

[0157] like Figure 3 As shown, the hospital smart fire monitoring system based on the Internet of Things provided by this application includes:

[0158] IoT device deployment module, used to obtain the spatial location coordinates of fire-fighting equipment, sensors, and surveillance cameras;

[0159] A prediction model training module is used to obtain N historical hospital interior images, annotate the fire category of each historical hospital interior image, wherein the fire category includes fire and no fire, and train a fire prediction model using the annotated historical hospital interior images;

[0160] The first-level warning trigger module is used to obtain images of the interior of the hospital, use the fire prediction model to predict the probability value of the hospital interior image belonging to each fire category, and determine whether there is a fire based on the probability value. If there is a fire, the first-level warning is triggered;

[0161] The first-level fire confirmation module is used to obtain environmental parameters and calculate the comprehensive fire risk value through sensors within n1 consecutive normal sampling intervals after the first-level warning is triggered. If the comprehensive fire risk value of n1 normal sampling intervals is less than or equal to the risk threshold, it is confirmed that there is no fire and the first-level warning is lifted;

[0162] The second-level fire confirmation module is used to confirm that there is a fire and trigger the second-level warning if the comprehensive fire risk value of n1 normal sampling intervals is greater than the risk threshold; calculate the adjustment coefficient according to the comprehensive fire risk value, obtain the second sampling interval according to the adjustment coefficient, and the sensor samples the abnormal environmental parameters according to the second sampling interval;

[0163] The fire extinguishing equipment startup module is used to obtain normal environmental parameters and calculate the difference value in combination with abnormal environmental parameters. If the difference value is greater than the difference threshold within n2 consecutive second sampling intervals, the spatial position coordinates of the sensor that obtains the abnormal environmental parameters are used as the fire point, and the N fire extinguishing equipment closest to the fire point are started.

[0164] Example 3

[0165] Figure 4 Schematic diagram of the structure of an electronic device provided by an embodiment of the present application. Figure 4 As shown, according to another aspect of the present application, an electronic device is also provided. The electronic device may include one or more processors and one or more memories. The memory stores a computer-readable code, and when the computer-readable code is executed by one or more processors, the hospital smart fire monitoring method based on the Internet of Things as described above can be executed.

[0166] The method or system according to the embodiment of the present application can also be used by Figure 4 The electronic device architecture shown in FIG. Figure 4As shown, the electronic device may include a bus, one or more CPUs, a read-only memory (ROM), a random access memory (RAM), a communication port connected to a network, an input / output component, a hard disk, and the like. A storage device in the electronic device, such as a ROM or a hard disk, may store the hospital smart fire monitoring method based on the Internet of Things provided in this application. The hospital smart fire monitoring method based on the Internet of Things may, for example, include: obtaining the spatial position coordinates of fire-fighting equipment, sensors, and surveillance cameras; obtaining N historical images of the interior of the hospital, marking the fire category of each historical image of the interior of the hospital, wherein the fire category includes fire and no fire, and using the labeled historical images of the interior of the hospital to train a fire prediction model; obtaining images of the interior of the hospital, and using the fire prediction model to predict the probability values ​​of the images of the interior of the hospital belonging to each fire category, and judging whether there is a fire based on the probability values. If there is a fire, a first-level warning is triggered; within n1 consecutive normal sampling intervals after the first-level warning is triggered, environmental parameters are obtained through sensors and a comprehensive fire risk value is calculated. If If the comprehensive fire risk value of n1 normal sampling intervals is less than or equal to the risk threshold, it is confirmed that there is no fire and the first-level warning is lifted; if the comprehensive fire risk value of n1 normal sampling intervals is greater than the risk threshold, it is confirmed that there is a fire and the second-level warning is triggered; the adjustment coefficient is calculated according to the comprehensive fire risk value, and the second sampling interval is obtained according to the adjustment coefficient. The sensor samples the abnormal environmental parameters according to the second sampling interval; the normal environmental parameters are obtained, and the difference value is calculated in combination with the abnormal environmental parameters. If the difference value is greater than the difference threshold within n2 consecutive second sampling intervals, the spatial position coordinates of the sensor that obtains the abnormal environmental parameters are used as the fire point, and the N fire-fighting devices closest to the fire point are activated. Furthermore, the electronic device may also include a user interface. Of course, Figure 4 The architecture shown is only exemplary and can be omitted according to actual needs when implementing different devices. Figure 4 One or more components of an electronic device are shown.

[0167] Example 4

[0168] Figure 5 Schematic diagram of a readable storage medium structure provided by an embodiment of the present application. Figure 5 As shown, it is a readable storage medium according to an embodiment of the present application. Computer-readable instructions are stored on the computer-readable storage medium. When the computer-readable instructions are executed by the processor, the hospital smart fire monitoring method based on the Internet of Things according to the embodiment of the present application described with reference to the above figures can be executed. The storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0169] In addition, according to the implementation mode of the present application, the process described in the above reference flowchart can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium, the non-transitory machine-readable storage medium stores machine-readable instructions, the machine-readable instructions can be executed by a processor to execute instructions corresponding to the method steps provided in the present application, for example: obtaining the spatial position coordinates of fire-fighting equipment, sensors, and surveillance cameras; obtaining N historical hospital interior images, marking the fire category of each historical hospital interior image, the fire category including fire and no fire, and using the marked historical hospital interior images to train a fire prediction model; obtaining hospital interior images, using the fire prediction model to predict the probability value of the hospital interior image belonging to each fire category, judging whether there is a fire based on the probability value, and if there is a fire, triggering the first level warning; after the first level warning is triggered, the continuous n Within 1 normal sampling interval, the environmental parameters are obtained through the sensor and the comprehensive fire risk value is calculated. If the comprehensive fire risk value of n1 normal sampling intervals is less than or equal to the risk threshold, it is confirmed that there is no fire and the first-level warning is lifted; if the comprehensive fire risk value of n1 normal sampling intervals is greater than the risk threshold, it is confirmed that there is a fire and the second-level warning is triggered; the adjustment coefficient is calculated according to the comprehensive fire risk value, and the second sampling interval is obtained according to the adjustment coefficient. The sensor samples the abnormal environmental parameters according to the second sampling interval; the normal environmental parameters are obtained, and the difference value is calculated in combination with the abnormal environmental parameters. If the difference value is greater than the difference threshold within n2 consecutive second sampling intervals, the spatial position coordinates of the sensor that obtains the abnormal environmental parameters are used as the fire point, and the N fire extinguishing devices closest to the fire point are activated. When the computer program is executed by the central processing unit (CPU), the above functions defined in the method of the present application are executed.

[0170] The methods, apparatuses, and devices of the present application may be implemented in many ways. For example, the methods, apparatuses, and devices of the present application may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above order of steps for the method is for illustration only, and the steps of the method of the present application are not limited to the order specifically described above unless otherwise specified. In addition, in some embodiments, the present application may also be implemented as programs recorded in a recording medium, which include machine-readable instructions for implementing the method according to the present application. Therefore, the present application also covers recording media storing programs for executing the method according to the present application.

[0171] In addition, the parts of the above-mentioned technical solutions provided in the embodiments of the present application that are consistent with the implementation principles of the corresponding technical solutions in the prior art are not described in detail to avoid excessive redundancy.

[0172] The specific implementation modes as described above further describe the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation mode of the present invention and is not intended to limit the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A hospital intelligent fire monitoring method based on the Internet of Things, characterized in that: include: Obtain the spatial location coordinates of fire-fighting equipment, sensors, and surveillance cameras; Obtain N historical hospital interior images, annotate the fire category of each historical hospital interior image, wherein the fire category includes fire and no fire, and use the annotated historical hospital interior images to train a fire prediction model; Obtain images of the interior of the hospital, use the fire prediction model to predict the probability values ​​of the hospital interior images belonging to each fire category, and determine whether there is a fire based on the probability values. If there is a fire, trigger the first level warning; Within n1 consecutive normal sampling intervals after the first-level warning is triggered, the environmental parameters are obtained through sensors and the comprehensive fire risk value is calculated. If the comprehensive fire risk value of n1 normal sampling intervals is less than or equal to the risk threshold, it is confirmed that there is no fire and the first-level warning is lifted; If the comprehensive fire risk value of n1 normal sampling intervals is greater than the risk threshold, it is confirmed that there is a fire and the second-level warning is triggered; An adjustment coefficient is calculated according to the comprehensive fire risk value, a second sampling interval is obtained according to the adjustment coefficient, and the sensor samples abnormal environmental parameters according to the second sampling interval; Normal environmental parameters are obtained and the difference value is calculated in combination with abnormal environmental parameters. If the difference value is greater than the difference threshold within n2 consecutive second sampling intervals, the spatial position coordinates of the sensor that obtains the abnormal environmental parameters are used as the fire point, and the N fire-fighting devices closest to the fire point are activated.

2. The hospital intelligent fire monitoring method based on the Internet of Things as claimed in claim 1, characterized in that: The specific method for obtaining the spatial position coordinates of the fire extinguishing equipment, sensors and surveillance cameras is: Deploy fire extinguishing equipment, a sensor network and a surveillance camera network inside the hospital, wherein the sensor network includes a temperature sensor, a smoke sensor and a toxic gas sensor; A sampling interval mechanism for sensors and surveillance cameras is set, and the sampling interval mechanism is switched according to the triggering and release of the first-level warning and the second-level warning. The sampling interval mechanism includes a normal sampling interval and a second sampling interval. The sensor network and the surveillance camera network use the normal sampling interval as the sampling interval mechanism in the absence of fire. When there is no fire and the first-level warning is triggered, the normal sampling interval is used as the sampling interval mechanism. When the second-level warning is triggered, the sampling interval mechanism is switched to the second sampling interval; A 3D model of the interior of the hospital was constructed, and a comprehensive scan of the interior of the hospital was performed using lidar to obtain point cloud data of fire-fighting equipment, sensor point cloud data, and camera point cloud data. The center point of each fire-fighting equipment point cloud, sensor point cloud, and camera point cloud was used as the spatial position coordinates of the corresponding fire-fighting equipment, sensor, and surveillance camera.

3. The hospital intelligent fire monitoring method based on the Internet of Things as claimed in claim 2 is characterized in that: The method of obtaining N historical hospital interior images, marking the fire category of each historical hospital interior image, wherein the fire category includes fire and no fire, and using the marked historical hospital interior images to train the fire prediction model is as follows: Obtain N historical hospital interior images from the hospital's surveillance fire database, wherein the historical hospital interior images include hospital interior images of different areas in the hospital and different historical time periods; Define the fire category label, 0 means no fire, 1 means fire, and label each historical hospital interior image with the fire category. If there is a fire feature, it is marked as 1, otherwise it is marked as 0; The annotated historical hospital internal images are used as training samples to construct the structure of a convolutional neural network model, which includes two convolutional layers, two maximum pooling layers, and two fully connected layers; The cross entropy loss function is used as the loss function, and the Adam optimizer is used as the optimizer. In each training round, the historical hospital internal images are used as input data, and the convolutional neural network model is used to predict the fire category of the historical hospital internal images, output the corresponding fire category label, calculate the loss function value, and update the model parameters according to the back propagation algorithm; The training objective is to minimize the loss function value between the predicted fire category label and the actual fire category label. When the loss function value converges, the training is completed and a trained fire prediction model is obtained.

4. The hospital intelligent fire monitoring method based on the Internet of Things as claimed in claim 3 is characterized in that: The specific method of obtaining the hospital interior image, using the fire prediction model to predict the probability value of the hospital interior image belonging to each fire category, judging whether there is a fire according to the probability value, and triggering the first level warning if there is a fire is as follows: A normal sampling interval f0 is set, where the normal sampling interval is the sampling interval of the surveillance camera and the sensor when there is no fire. During each normal sampling interval, the surveillance camera inside the hospital performs real-time sampling to obtain an image of the inside of the hospital; The hospital internal image is used as input data, and the fire prediction model is used to predict the probability value of the hospital internal image belonging to the two fire categories of fire and no fire. If the probability value of the output fire category is less than p1%, the normal sampling interval is maintained to sample the surveillance camera and sensor; If the output fire category has a probability value of fire greater than or equal to p1%, the first level warning is triggered.

5. The hospital intelligent fire monitoring method based on the Internet of Things as claimed in claim 4 is characterized in that: The specific method of lifting the first-level warning is as follows: for the hospital internal image predicted to have a fire, find the surveillance camera that obtains the hospital internal image, and use the spatial position coordinates of the surveillance camera as the fire shooting point; Calculate the distance between the fire shooting point (x, y, z) and the spatial position coordinates of each sensor, and find the temperature sensor, smoke sensor and gas sensor closest to the fire shooting point; Set the number of continuous sampling times n1. Within the n1 normal sampling intervals after the first-level warning is triggered, the temperature sensor, smoke sensor and toxic gas sensor are used to sample and obtain environmental parameters. The environmental parameters include ambient temperature, smoke concentration and toxic gas concentration. The local temperature distribution T = {T1, T2, ..., T n1 }、Local smoke concentration distribution S={S1,S2,...,S n1 }、Local gas concentration distribution G={G1,G2,...,G n1 }, where T n1 , S n1 , G n1 They represent the ambient temperature, smoke concentration, and toxic gas concentration of the n1th normal sampling interval respectively; Within n1 normal sampling intervals after the first-level warning is triggered, the comprehensive fire risk value of the environmental parameters sampled in the normal sampling interval is calculated; Preset risk threshold H th , compare the comprehensive fire risk value within n1 normal sampling intervals with the risk threshold. If the comprehensive fire risk value within n1 normal sampling intervals is less than or equal to the risk threshold, it is confirmed that there is no fire and the first level warning is lifted.

6. The hospital intelligent fire monitoring method based on the Internet of Things as claimed in claim 5 is characterized in that: The calculation method of the comprehensive fire risk value is: According to the local temperature distribution, local smoke concentration distribution and local toxic gas concentration distribution, calculate the average ambient temperature T within n1 normal sampling intervals. avg , average smoke density S avg and the average toxic gas concentration G avg ; Get normal environmental parameters, which include: normal ambient temperature T0, normal smoke concentration S0 and normal toxic gas concentration G0, combined with the average ambient temperature T avg , average smoke density S avg and the average toxic gas concentration G avg Calculate the temperature risk value R of the i-th normal sampling interval T,i , smoke risk value R S,i and the toxic gas risk value R G,i ; The fire risk value H of the ith normal sampling interval is calculated from the temperature risk value, smoke risk value and toxic gas risk value of the ith normal sampling interval. i ; According to the fire risk value of each normal sampling interval within n1 normal sampling intervals, the comprehensive fire risk value H of n1 normal sampling intervals is calculated. total .

7. The hospital intelligent fire monitoring method based on the Internet of Things as claimed in claim 6 is characterized in that: The specific method of calculating the adjustment coefficient according to the comprehensive fire risk value, obtaining the second sampling interval according to the adjustment coefficient, and the sensor sampling the abnormal environment parameter according to the second sampling interval is: Set the range of adjustment coefficient k to get the minimum adjustment coefficient k min and the maximum adjustment factor k max ; Calculate the difference k between the maximum adjustment coefficient and the minimum adjustment coefficient max -k min Calculate the risk of the comprehensive fire risk value compared to the risk threshold value H total -H th , divide the excess risk by the risk threshold to get the proportionality coefficient The proportional coefficient is multiplied by the difference to obtain the adjustment amount, and the adjustment amount is added to the minimum adjustment coefficient to obtain the adjustment coefficient k; Multiply the adjustment coefficient k by the normal sampling interval f0 to obtain the second sampling interval f2; The sensor samples abnormal environmental parameters according to the second sampling interval, wherein the abnormal environmental parameters include abnormal environmental temperature T yc , Abnormal smoke density S yc , Abnormal toxic gas concentration G yc .

8. A hospital smart fire monitoring system based on the Internet of Things, which is implemented based on the hospital smart fire monitoring method based on the Internet of Things as claimed in any one of claims 1 to 7, characterized in that: include: IoT device deployment module, used to obtain the spatial location coordinates of fire-fighting equipment, sensors, and surveillance cameras; A prediction model training module is used to obtain N historical hospital interior images, annotate the fire category of each historical hospital interior image, wherein the fire category includes fire and no fire, and train a fire prediction model using the annotated historical hospital interior images; The first-level warning trigger module is used to obtain images of the interior of the hospital, use the fire prediction model to predict the probability value of the hospital interior image belonging to each fire category, and determine whether there is a fire based on the probability value. If there is a fire, the first-level warning is triggered; The first-level fire confirmation module is used to obtain environmental parameters and calculate the comprehensive fire risk value through sensors within n1 consecutive normal sampling intervals after the first-level warning is triggered. If the comprehensive fire risk value of n1 normal sampling intervals is less than or equal to the risk threshold, it is confirmed that there is no fire and the first-level warning is lifted; The second-level fire confirmation module is used to confirm that there is a fire and trigger the second-level warning if the comprehensive fire risk value of n1 normal sampling intervals is greater than the risk threshold; An adjustment coefficient is calculated according to the comprehensive fire risk value, a second sampling interval is obtained according to the adjustment coefficient, and the sensor samples abnormal environmental parameters according to the second sampling interval; The fire extinguishing equipment startup module is used to obtain normal environmental parameters and calculate the difference value in combination with abnormal environmental parameters. If the difference value is greater than the difference threshold within n2 consecutive second sampling intervals, the spatial position coordinates of the sensor that obtains the abnormal environmental parameters are used as the fire point, and the N fire extinguishing equipment closest to the fire point are started.

9. An electronic device, characterized in that: It includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps in the hospital smart fire monitoring method based on the Internet of Things are implemented as described in any one of claims 1 to 7.

10. A readable storage medium, characterized in that: The readable storage medium stores a computer program, which is suitable for being loaded by a processor to execute the steps in the hospital smart fire monitoring method based on the Internet of Things as described in any one of claims 1-7.

Images