A high-rise fire intelligent positioning voice alarm method and system

Abstract

The invention discloses a high-rise fire intelligent positioning voice alarm method and system, which specifically relates to the field of voice alarm technology, including taking the concentration of different types of smoke generated by the combustion of different materials, the brightness value of the flame and the spread speed of the flame as the input of a fuzzy Bayesian neural network model, taking the probability distribution of the fire hazard level as the output of the fuzzy Bayesian neural network model, determining the fire hazard score of the fire area according to the probability distribution of the fire hazard level, collecting the crowd flow information and the regional geographical location information in the monitoring area, comprehensively analyzing the crowd flow information and the geographical location information in the monitoring area, quantifying the adjustment requirements of the alarms in each monitoring area in the building, and generating the adjustment signals of the alarms in real time. The invention improves the efficiency and accuracy of fire monitoring, combines the specific situation of the fire with the alarm level, and realizes the flexibility and effectiveness of the fire emergency response.

Description

Technical Field

[0001] The present invention relates to the field of voice alarm technology, and more specifically, to a high-rise fire intelligent positioning voice alarm method and system. Background Art

[0002] Fire is a sudden disaster that is usually accompanied by the generation of smoke and flames, posing a serious threat to human life and property safety. Modern buildings are usually equipped with fire alarm systems that can detect fires and sound alarms in a timely manner through smoke sensors and heat sensors. However, the degree of fire danger varies on different floors and areas. A single alarm mode cannot effectively adapt to the actual risk situation in different environments, and the degree of fire danger is not easily linked to the alarm emergency level.

[0003] In order to solve the above defects, a technical solution is now provided. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present invention provide a high-rise fire intelligent positioning voice alarm method and system to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A high-rise fire intelligent positioning voice alarm method, characterized in that it specifically includes the following steps:

[0007] S1: The concentration of different types of smoke generated by the combustion of different materials, the brightness value of the flame and the spread speed of the flame are used as the input of the fuzzy Bayesian neural network model, and the probability distribution of the fire hazard level is used as the output of the fuzzy Bayesian neural network model;

[0008] S2: Determine the fire hazard score of the fire area based on the probability distribution of the fire hazard level, and collect crowd flow information and regional geographical location information in the monitoring area based on the geographical location relationship between the fire area and the monitoring area;

[0009] S3: Comprehensively analyze the crowd flow information and geographic location information in the monitoring area to quantify the adjustment requirements of the alarms in each monitoring area in the building;

[0010] S4: Based on the comprehensive analysis results of the crowd flow information and regional geographical location information in the monitoring area at different time periods, an adjustment signal for the alarm is generated in real time;

[0011] Among them, the crowd flow information in the monitoring area includes:

[0012] The crowd flow information is expressed by the crowd abnormal aggregation coefficient and the unit time period is set;

[0013] The logic for obtaining the abnormal crowd aggregation coefficient is as follows: the flow of people in the area is captured by a camera, the number of people at different times in a unit time period is obtained, and the number of people at different times in a unit time period is marked as: RQ h , where h = 1, 2, 3, ..., H, H is a positive integer, and h is the number of different moments;

[0014] Set the crowd quantity threshold and mark it as: RQ yz , compare the number of people at different times in a unit time period with the crowd number threshold, obtain the time when the number of people is greater than the crowd number threshold, and re-mark the time when the number of people at different times in a unit time period is greater than the crowd number threshold as: RQ j , where j=1, 2, 3, ..., J, J is a positive integer, and j is the number of the moment when the number of people at different moments in a unit time period is greater than the threshold of the number of people;

[0015] Calculate the abnormal clustering coefficient of the crowd, the calculation formula is: Among them, YC rq is the abnormal clustering coefficient of the crowd;

[0016] Geographical location information of the monitored area, including:

[0017] The regional geographical location information is represented by a safety position deviation coefficient;

[0018] The logic for obtaining the safety position deviation coefficient is as follows: obtain the floor where the fire source is located, the floor where the monitoring area is located, and the distance between the monitoring area and the safe evacuation area, and obtain the floor deviation value. The calculation formula of the floor deviation value is: Among them, LC pc is the floor deviation value, HY LC is the floor where the fire originated, QY LC is the floor where the monitoring area is located, R is the correction factor, ensuring that the denominator is a positive integer, which is used to adjust the impact of the fire source floor on safety;

[0019] Obtain the location coordinates of the monitoring area and the safe evacuation area, and calculate the safe evacuation deviation value. The calculation formula for the safe evacuation deviation value is: Among them, CL pc is the safe evacuation deviation value, (x jc ,y jc ) is the location coordinate of the monitoring area, (x cl ,y cl ) are the location coordinates of the safe evacuation area;

[0020] Calculate the safe position deviation coefficient using the following formula: Among them, PCaq is the safety position deviation coefficient.

[0021] In a preferred embodiment, the concentration of different types of smoke generated by the combustion of different materials, the brightness value of the flame, and the spread speed of the flame are used as inputs of the fuzzy Bayesian neural network model, including:

[0022] Specific gases are identified through specific sensors, monitoring intervals are set, and the average concentrations of different types of smoke in the monitoring intervals are marked as: ND n , where n = 1, 2, 3, ..., N, N is a positive integer, and n is the number of the smoke type;

[0023] The brightness of the flame is captured by optical sensors, cameras and other equipment, and the average brightness value of the flame in the monitoring interval is marked as: LD;

[0024] The size and position of each flame frame are captured by the camera to determine the flame spread speed information, which is then expressed by the fire spread accumulation coefficient;

[0025] The logic for obtaining the fire spread accumulation coefficient is as follows: obtain the initially captured flame image, extract the number of pixels in the flame area in the initial flame image, and mark the number of pixels in the flame area in the initial flame image as: XL cs ;

[0026] Obtaining each frame of the flame image within the monitoring interval, and extracting the number of pixels belonging to the flame area in each frame of the flame image within the monitoring interval, determining the variation function of the number of pixels in the flame area within the monitoring interval and time according to the time length of the monitoring interval, and marking the variation function of the number of pixels in the flame area within the monitoring interval and time as: XL(t), wherein the variation function of the number of pixels and time is determined by linear or nonlinear fitting;

[0027] The monitoring interval is marked as: T, and the monitoring interval is divided into the first half and the second half, and the duration of each half is Calculate the fire spread accumulation coefficient using the following formula: Among them, MY jl is the fire spread accumulation coefficient, MY(t)=XL(t)-XL cs MY(t) is the difference function between the number of pixels in the flame area in the monitoring interval and the time and the number of pixels in the flame area in the initial flame image, t a ~t b is the time period in the first half of the monitoring interval that is smaller than the number of pixels in the flame area in the initial flame image, t c ~t dIt is the time period in the second half of the monitoring interval where the number of pixels in the flame area is greater than that in the initial flame image.

[0028] In a preferred embodiment, determining the fire risk score of the fire area includes:

[0029] The output parameter of the fuzzy Bayesian neural network model is the probability distribution of fire hazard level. The fire hazard level is classified into different levels, including low hazard level, medium hazard level and high hazard level. According to the training results of the fuzzy Bayesian neural network model, the probability distribution of fire hazard level is determined, and the fire hazard score of the fire area is determined;

[0030] According to the probability distribution of fire hazard level, the fire hazard score is determined by weighted summation calculation. The calculation formula of fire hazard score is: PF hz =α1y d +α2y z +α3y g PF hz Score the fire hazard, y d Low risk level, y z For medium risk level, y g is a high risk level, α1, α2, and α3 are the proportional coefficients of low risk level, medium risk level, and high risk level, respectively, and α1, α2, and α3 are all greater than 0.

[0031] In a preferred embodiment, quantifying the adjustment requirements of alarms in each monitoring area in a building includes:

[0032] According to the fire hazard score of the fire, the abnormal crowd aggregation coefficient and the safety position deviation coefficient are weighted and calculated to build an alarm adjustment evaluation model and generate an alarm adjustment evaluation coefficient. The expression of the alarm adjustment evaluation coefficient is: Among them, pg tj is the alarm adjustment evaluation coefficient, β1 and β2 are the proportional coefficients of the abnormal crowd aggregation coefficient and the safe position deviation coefficient, respectively, and β1 and β2 are both greater than 0.

[0033] In a preferred embodiment, generating an adjustment signal of an alarm in real time includes:

[0034] Obtain the alarm adjustment evaluation coefficient of each monitoring area in real time, adjust the alarm of the monitoring area according to the alarm adjustment evaluation coefficient of each unit time period of the monitoring area, and compare the alarm adjustment evaluation coefficient of the previous unit time period of the monitoring area with the alarm adjustment evaluation coefficient of the current unit time period;

[0035] If the alarm adjustment evaluation coefficient of the previous unit time period of the monitoring area is greater than the alarm adjustment evaluation coefficient of the current unit time period, a reduction signal is generated;

[0036] If the alarm adjustment evaluation coefficient of the previous unit time period of the monitoring area is less than the alarm adjustment evaluation coefficient of the current unit time period, an increase signal is generated.

[0037] In a preferred embodiment, a high-rise fire intelligent positioning voice alarm system includes a data acquisition module, a neural network module, an evaluation module and a signal generation module, and the modules are signal-connected;

[0038] Data collection module, used to collect real-time information on different types of smoke concentrations, flame brightness values, flame spread speed information, crowd flow information in the monitored area, and regional geographic location information of the monitored area;

[0039] The neural network module is used to construct a fuzzy Bayesian neural network model to output the probability distribution of the fire hazard level according to the different types of smoke concentration, flame brightness value, and flame spread speed information;

[0040] An evaluation module is used to construct an alarm adjustment evaluation model based on the crowd flow information and the regional geographical location information of the monitoring area, generate an alarm adjustment evaluation coefficient, and quantify the adjustment requirements of the alarm;

[0041] The signal generation module is used to adjust the evaluation coefficient of the alarm in different time periods, generate a reduction signal or an increase signal, and realize intelligent adjustment.

[0042] Technical effects and advantages of the present invention:

[0043] The present invention determines the fire situation in the fire area through smoke gas concentration sensors and visual monitoring equipment, uses a fuzzy Bayesian neural network model to quantify the danger level of the fire area, and connects the alarms in different areas of each floor through a central control center. In combination with the geographical location of the alarms, intelligent alarms are realized for different areas on different floors. The present invention improves the efficiency and accuracy of fire monitoring, combines the specific situation of the fire with the alarm level, and realizes the flexibility and effectiveness of fire emergency response. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to facilitate understanding by those skilled in the art, the present invention is further described below in conjunction with the accompanying drawings;

[0045] Figure 1 A schematic diagram of a flow chart of a high-rise fire intelligent positioning voice alarm method of the present invention;

[0046] Figure 2 The present invention is a structural schematic diagram of a high-rise fire intelligent positioning voice alarm system. DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] Example 1

[0049] like Figure 1 A flow chart of a high-rise fire intelligent positioning voice alarm method of the present invention is provided, which specifically comprises the following steps:

[0050] S1: The concentration of different types of smoke generated by the combustion of different materials, the brightness value of the flame and the spread speed of the flame are used as the input of the fuzzy Bayesian neural network model, and the probability distribution of the fire hazard level is used as the output of the fuzzy Bayesian neural network model;

[0051] S2: Determine the fire hazard score of the fire area based on the probability distribution of the fire hazard level, and collect crowd flow information and regional geographical location information in the monitoring area based on the geographical location relationship between the fire area and the monitoring area;

[0052] S3: Comprehensively analyze the crowd flow information and geographic location information in the monitoring area to quantify the adjustment requirements of the alarms in each monitoring area in the building;

[0053] S4: Based on the comprehensive analysis results of the crowd flow information and regional geographical location information in the monitoring area at different time periods, an adjustment signal for the alarm is generated in real time.

[0054] In step 1, the fuzzy Bayesian neural network model is trained by collecting sample data, and the input parameters of the fuzzy Bayesian neural network model are the average concentration of different types of smoke, the average brightness value of the flame, and the fire spread accumulation coefficient;

[0055] Specifically, a specific gas is identified through a specific sensor, a monitoring interval is set, and the average concentration of different types of smoke in the monitoring interval is marked as: ND n , where n = 1, 2, 3, ..., N, N is a positive integer, and n is the number of the smoke type;

[0056] It should be noted that the monitoring interval is a specific time period. By collecting the smoke concentration at different times in the monitoring interval through sensors, the average smoke concentration level in the monitoring interval can be obtained, which can more comprehensively reflect the overall situation of smoke and reduce the inaccuracy caused by sensor recognition errors.

[0057] The combustion of different materials will produce different types of smoke, including carbon dioxide, carbon monoxide, etc. The concentration of different types of smoke has different degrees of harm to the human body, and generally the higher the concentration of any type of smoke, the higher the degree of harm to the human body. The concentration of smoke and the degree of harm to the human body are determined through professional evaluation and standards of fire protection.

[0058] The brightness of the flame is captured by optical sensors, cameras and other equipment, and the average brightness value of the flame in the monitoring interval is marked as: LD;

[0059] It should be noted that the brightness of a flame is directly related to its intensity and heat output. Generally, the brighter the flame, the more complete its combustion, the higher the temperature, and the greater the heat and energy released. By analyzing the brightness of the flame, the efficiency of the combustion can be evaluated. There is also a certain relationship between the brightness of the flame and the smoke concentration. An increase in smoke concentration may cause the brightness of the flame to decrease. Therefore, the combination of flame brightness and smoke concentration can provide a more comprehensive evaluation.

[0060] The size and position of each flame frame are captured by the camera to determine the flame spread speed information, which is then expressed by the fire spread accumulation coefficient;

[0061] The size and position of each flame frame are captured by the camera, including:

[0062] Use a camera to capture flame images in real time and record flame changes at a certain frame rate (e.g., 30 frames per second);

[0063] Perform noise processing and color enhancement on each frame to ensure the clarity of the flame area, and perform color threshold segmentation on the image or use a convolutional neural network (CNN) to extract the flame area so that the flame area stands out from the background;

[0064] In the preprocessed image, the flame area is binarized so that the flame part is displayed as white (pixel value 1) and the background is black (pixel value 0), and the number of white pixels in the flame area is counted.

[0065] The logic for obtaining the fire spread accumulation coefficient is as follows: obtain the initially captured flame image, extract the number of pixels in the flame area in the initial flame image, and mark the number of pixels in the flame area in the initial flame image as: XL cs ;

[0066] It should be noted that the flame image is an image captured by a camera. The image has a fixed format and a fixed size. The number of pixels belonging to the flame area is extracted through an image recognition algorithm. The more pixels belonging to the flame area, the stronger the flame.

[0067] Obtaining each frame of the flame image within the monitoring interval, and extracting the number of pixels belonging to the flame area in each frame of the flame image within the monitoring interval, determining the variation function of the number of pixels in the flame area within the monitoring interval and time according to the time length of the monitoring interval, and marking the variation function of the number of pixels in the flame area within the monitoring interval and time as: XL(t), wherein the variation function of the number of pixels and time is determined by linear or nonlinear fitting;

[0068] The monitoring interval is marked as: T, and the monitoring interval is divided into the first half and the second half, and the duration of each half is Calculate the fire spread accumulation coefficient using the following formula: Among them, MY jl is the fire spread accumulation coefficient, MY(t)=XL(t)-XL cs MY(t) is the difference function between the number of pixels in the flame area in the monitoring interval and the time and the number of pixels in the flame area in the initial flame image, t a ~t b is the time period in the first half of the monitoring interval that is smaller than the number of pixels in the flame area in the initial flame image, t c ~t d It is the time period in the second half of the monitoring interval where the number of pixels in the flame area is greater than that in the initial flame image.

[0069] It can be seen from the formula that the larger the fire spread accumulation coefficient, the higher the possibility of fire spread, which means the faster the fire expands and the higher the degree of danger.

[0070] The output parameter of the fuzzy Bayesian neural network model is the probability distribution of fire hazard level. The fire hazard level is classified into different levels, including low hazard level, medium hazard level and high hazard level. The fuzzy Bayesian neural network allows labels to exist in the form of probability distribution. For example: high hazard level sample: [0.0, 0.1, 0.9] means that most of the probabilities are "high hazard". Ensure that the output probability is calibrated. Use calibration techniques such as temperature scaling to optimize the model output so that the predicted probability is consistent with the actual frequency of occurrence.

[0071] The fuzzy Bayesian neural network model is trained according to the sample data, and the sample training set of the Bayesian neural network model is obtained by normalizing the sample data, that is, in, is the average smoke concentration of the nth smoke type in the mth sample data, m = 1, 2, 3, ..., i, i is a positive integer, i represents the number of sample data collected, m is the number of sample data, LD m is the average brightness value of the flame of the mth sample data, is the fire spread accumulation coefficient of the mth sample data, is the probability distribution of fire hazard level of the mth sample data.

[0072] In the fuzzy Bayesian neural network model, a prior distribution is assigned to each weight, and a fuzzy prior distribution is given to the weight. It is assumed that the prior distribution of each weight is a normal distribution. The standard deviation σ is regarded as a fuzzy variable, and fuzziness is introduced into the standard deviation σ in the normal distribution prior distribution of each weight, and a fuzzy range is set. in, is a fuzzy number indicating a higher or lower confidence in the model's weights.

[0073] In each training iteration, the loss function is calculated and the posterior distribution of the weights is updated based on the data. Due to the existence of fuzzy uncertainty, the fuzzy Bayesian inference method is used to update the posterior distribution of each weight and adjust each weight in the fuzzy interval. Confidence within.

[0074] Define the structure of the fuzzy Bayesian neural network model, including input layer, hidden layer, and output layer. The input features of the input layer are the average concentration of different types of smoke, the average brightness value of the flame, and the accumulation coefficient of fire spread. Therefore, the number of neuron nodes in the input layer is N+2, and the number of neuron nodes in the hidden layer is determined by an empirical formula. The output feature of the output layer is the probability distribution of the fire hazard level. The number of neuron nodes is 3, and the empirical formula is: Among them, G is the number of neuron nodes in the hidden layer, h is the number of neuron nodes in the input layer, m is the number of neuron nodes in the output layer, and a is a constant ranging from 1 to 10.

[0075] By randomly sampling weight values ​​from the current prior distribution, the sampled weights are used to calculate the predicted output of the model. According to the difference between the predicted result and the actual label, the loss function is calculated. The gradient of the loss function with respect to the weight is calculated through optimization algorithms such as gradient descent. The weight distribution is updated according to the loss to gradually converge to the appropriate posterior distribution. The convergence of the training process is determined by monitoring the changes in the loss function. If the loss drops to a certain threshold or reaches the preset number of iterations, the training is stopped.

[0076] In step 2, the probability distribution of the fire hazard level is determined according to the training results of the fuzzy Bayesian neural network model. According to the probability distribution of the fire hazard level, the fire hazard score is determined by weighted summation calculation. The calculation formula of the fire hazard score is: PF hz =α1y d +α2y z +α3y g PF hzis the fire hazard score, α1, α2, and α3 are the proportional coefficients of low hazard level, medium hazard level, and high hazard level, respectively, and α1, α2, and α3 are all greater than 0.

[0077] It should be noted that, for example, at a high level of fire hazard score, the voice alarm will be louder, the alarm frequency will be higher, and the voice alarm pattern may be different in different areas of a high-rise building. The higher the fire hazard score, the louder the voice alarm may be in an area far away from the fire. Conversely, the lower the fire hazard score, the smaller the voice alarm may be in an area far away from the fire.

[0078] By identifying the location of the fire, the fire location is sent to the central control center, which connects the alarms in different areas of each floor to realize intelligent alarms in different areas of different floors, and combines the geographical location data of the alarm to realize an intelligent and regionalized fire alarm system, including:

[0079] Sensor network: Install fire detectors such as smoke, temperature, and flame sensors in different areas of each floor to monitor fire signals in real time;

[0080] Location identification: When a fire signal is detected in a certain area, the sensor will send the location information including the floor, specific area and fire parameters to the central control center;

[0081] Data transmission: Use wireless or wired IoT transmission protocols (such as LoRa, Zigbee, Wi-Fi, etc.) to efficiently transmit real-time information and location information of fire to the central control center;

[0082] Intelligent regional alarm: Through floor zoning maps and regional IDs, the alarms in the fire area and adjacent areas are accurately activated to avoid indiscriminate alarms throughout the building. The volume and frequency of the alarm voice broadcast are adjusted to reduce panic and guide people in the areas with the greatest fire risk to evacuate first.

[0083] Alarm geographic location data: The specific location and floor location of each alarm is entered into the central control system in advance. The control center sends voice evacuation instructions based on the fire location and recommends safe escape routes based on the floor map and fire location.

[0084] After determining the fire hazard score according to the fuzzy Bayesian neural network model, the crowd flow information and regional geographical location information of each high-level monitoring area are collected in combination with the monitoring equipment. After collection, the crowd flow information is represented by the crowd abnormal aggregation coefficient, and the regional geographical location information is represented by the safe position deviation coefficient. The crowd flow information and the regional geographical location information are comprehensively analyzed to construct an alarm adjustment evaluation model. Through the alarm adjustment evaluation model, the alarms in each area are adjusted in real time.

[0085] Set a unit time period, and use monitoring equipment to monitor the crowd flow information and regional geographical location information of each area within the unit time period, where the unit time period is a specific and fixed length of time set by professional staff. The monitoring equipment in each area collects data within the unit time period and analyzes it to obtain the crowd flow information and regional geographical location information.

[0086] The logic for obtaining the abnormal crowd aggregation coefficient is as follows: the flow of people in the area is captured by a camera, the number of people at different times in a unit time period is obtained, and the number of people at different times in a unit time period is marked as: RQ h , where h = 1, 2, 3, ..., H, H is a positive integer, and h is the number of different moments;

[0087] It should be noted that computer vision technology is used for crowd detection and tracking, and deep learning models are used to detect the number of people in the video image at each moment, such as YOLO, Faster R-CNN and other computer vision technologies.

[0088] Set the crowd quantity threshold and mark it as: RQ yz , compare the number of people at different times in a unit time period with the crowd number threshold, obtain the time when the number of people is greater than the crowd number threshold, and re-mark the time when the number of people at different times in a unit time period is greater than the crowd number threshold as: RQ j , where j=1, 2, 3, ..., J, J is a positive integer, and j is the number of the moment when the number of people at different moments in a unit time period is greater than the threshold of the number of people;

[0089] It should be noted that the crowd size threshold is set by professional staff and is determined by the size of the camera monitoring area and the upper limit of the number of people that can be accommodated in the monitoring area. Usually, if the crowd size threshold is exceeded in the monitoring area, the possibility of unsafe factors such as crowd pushing and trampling will be higher.

[0090] Calculate the abnormal clustering coefficient of the crowd, the calculation formula is: Among them, YC rq is the abnormal clustering coefficient of the crowd.

[0091] It can be seen from the formula that the larger the abnormal crowd aggregation coefficient is, the more abnormal crowd aggregation there is in a unit time period, indicating that the fire causes people in the area to gather and evacuate quickly, and there are certain safety hazards. The abnormal crowd aggregation coefficient can be fed back to the central control center, which will adjust the urgency of the alarm in the area to reduce people's panic.

[0092] Among them, the central control center can adjust the urgency of the sirens in the area from the following aspects to match the flow of people and reduce congestion and panic, including:

[0093] Volume: When the density of people is high, the alarm volume is lowered to prevent excessive tension and reduce congestion; when the density of people is low or during the evacuation process, the volume is gradually increased to remind people who have not yet evacuated;

[0094] Frequency: Increase or decrease the frequency of alarms. High-frequency alarms can increase people's vigilance, while low-frequency alarms can reduce tension and facilitate gradual evacuation.

[0095] Flashing frequency: By adjusting the flashing frequency of the light, the evacuation speed of people is guided. High-frequency flashing indicates a higher urgency and encourages quick evacuation; low-frequency flashing indicates that the situation is relatively stable and helps reduce congestion.

[0096] The logic for obtaining the safety position deviation coefficient is as follows: obtain the floor where the fire source is located, the floor where the monitoring area is located, and the distance between the monitoring area and the safe evacuation area, and obtain the floor deviation value. The calculation formula of the floor deviation value is: Among them, LC pc is the floor deviation value, HY LC is the floor where the fire originated, QY LC is the floor where the monitoring area is located, R is the correction factor, ensuring that the denominator is a positive integer, which is used to adjust the impact of the fire source floor on safety;

[0097] It should be noted that the heat and smoke generated by the fire will spread upward rapidly, and the higher the floor above the fire source, the longer the evacuation time will be, and the number of trapped people may be higher. Therefore, the risk on the floors above the fire source may be higher, and the closer the floor is to the fire source, the higher the risk may be.

[0098] Obtain the location coordinates of the monitoring area and the safe evacuation area, and calculate the safe evacuation deviation value. The calculation formula for the safe evacuation deviation value is: Among them, CL pc is the safe evacuation deviation value, (xjc, yjc) is the position coordinate of the monitoring area, and (xcl, ycl) is the position coordinate of the safe evacuation area;

[0099] It should be noted that the location of the monitoring area and the safe evacuation area are fixed. The location of the monitoring area can be determined based on the location of the monitoring equipment. The safe evacuation area is usually the emergency exit, which is also a set area. There may be multiple monitoring areas and multiple safe evacuation areas on the same floor, but they are all located on the same floor.

[0100] Calculate the safe position deviation coefficient using the following formula: Among them, PC aq is the safety position deviation coefficient.

[0101] It can be seen from the formula that the larger the safety position deviation coefficient is, the worse the location of the monitored area is, and it is more difficult for people in the monitored area to evacuate, which may pose a greater risk. The safety position deviation coefficient can be fed back to the central control center, which will adjust the urgency of the alarm in the area and remind people in the area to evacuate quickly by increasing the volume and frequency of the alarm.

[0102] In step 3, according to the fire hazard score of the fire, the abnormal crowd aggregation coefficient and the safety position deviation coefficient are weighted and calculated to build an alarm adjustment evaluation model and generate an alarm adjustment evaluation coefficient. The expression of the alarm adjustment evaluation coefficient is: Among them, pg tj is the alarm adjustment evaluation coefficient, β1 and β2 are the proportional coefficients of the abnormal crowd aggregation coefficient and the safe position deviation coefficient, respectively, and β1 and β2 are both greater than 0.

[0103] It can be seen from the formula that the smaller the abnormal crowd aggregation coefficient and the smaller the safety position deviation coefficient, the smaller the alarm adjustment evaluation coefficient, which means that the monitoring area needs to reduce the urgency of the alarm in the area. Conversely, the smaller the abnormal crowd aggregation coefficient and the larger the safety position deviation coefficient, the larger the alarm adjustment evaluation coefficient, which means that the monitoring area needs to increase the urgency of the alarm in the area.

[0104] In step 4, the alarm adjustment evaluation coefficient of each monitoring area is obtained in real time, and the alarm of the monitoring area is adjusted according to the alarm adjustment evaluation coefficient of each unit time period of the monitoring area. The alarm adjustment evaluation coefficient of the previous unit time period of the monitoring area is compared with the alarm adjustment evaluation coefficient of the current unit time period. If the alarm adjustment evaluation coefficient of the previous unit time period of the monitoring area is greater than the alarm adjustment evaluation coefficient of the current unit time period, a reduction signal is generated, i.e., the urgency of the alarm in the monitoring area is reduced. If the alarm adjustment evaluation coefficient of the previous unit time period of the monitoring area is less than the alarm adjustment evaluation coefficient of the current unit time period, an increase signal is generated, i.e., the urgency of the alarm in the monitoring area is increased.

[0105] The present invention determines the fire situation in the fire area through smoke gas concentration sensors and visual monitoring equipment, uses a fuzzy Bayesian neural network model to quantify the danger level of the fire area, and connects the alarms in different areas of each floor through a central control center. In combination with the geographical location of the alarms, intelligent alarms are realized for different areas on different floors. The present invention improves the efficiency and accuracy of fire monitoring, combines the specific situation of the fire with the alarm level, and realizes the flexibility and effectiveness of fire emergency response.

[0106] Example 2

[0107] like Figure 2 A structural schematic diagram of a high-rise fire intelligent positioning voice alarm system of the present invention is given, which specifically includes a data acquisition module, a neural network module, an evaluation module and a signal generation module, and the modules are connected by signals;

[0108] Data collection module, used to collect real-time information on different types of smoke concentrations, flame brightness values, flame spread speed information, crowd flow information in the monitored area, and regional geographic location information of the monitored area;

[0109] The neural network module is used to construct a fuzzy Bayesian neural network model to output the probability distribution of the fire hazard level according to the different types of smoke concentration, flame brightness value, and flame spread speed information;

[0110] An evaluation module is used to construct an alarm adjustment evaluation model based on the crowd flow information and the regional geographical location information of the monitoring area, generate an alarm adjustment evaluation coefficient, and quantify the adjustment requirements of the alarm;

[0111] The signal generation module is used to adjust the evaluation coefficient of the alarm in different time periods, generate a reduction signal or an increase signal, and realize intelligent adjustment.

[0112] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0113] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0114] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0115] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0116] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0117] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0118] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage media include: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks or optical disks.

[0119] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A high-rise fire intelligent positioning voice alarm method, characterized in that: The specific steps include: S1: The concentration of different types of smoke generated by the combustion of different materials, the brightness value of the flame and the spread speed of the flame are used as the input of the fuzzy Bayesian neural network model, and the probability distribution of the fire hazard level is used as the output of the fuzzy Bayesian neural network model; S2: Determine the fire hazard score of the fire area based on the probability distribution of the fire hazard level, and collect crowd flow information and regional geographical location information in the monitoring area based on the geographical location relationship between the fire area and the monitoring area; S3: Comprehensively analyze the crowd flow information and geographic location information in the monitoring area to quantify the adjustment requirements of the alarms in each monitoring area in the building; S4: Based on the comprehensive analysis results of the crowd flow information and regional geographical location information in the monitoring area at different time periods, an adjustment signal for the alarm is generated in real time; Among them, the crowd flow information in the monitoring area includes: The crowd flow information is expressed by the crowd abnormal aggregation coefficient and the unit time period is set; The logic for obtaining the abnormal crowd aggregation coefficient is as follows: the flow of people in the area is captured by a camera, the number of people at different times in a unit time period is obtained, and the number of people at different times in a unit time period is marked as: RQ h , where h = 1, 2, 3, ..., H, H is a positive integer, and h is the number of different moments; Set the crowd quantity threshold and mark it as: RQ yz , compare the number of people at different times in a unit time period with the crowd number threshold, obtain the time when the number of people is greater than the crowd number threshold, and re-mark the time when the number of people at different times in a unit time period is greater than the crowd number threshold as: RQ j , where j=1, 2, 3, ..., J, J is a positive integer, and j is the number of the moment when the number of people at different moments in a unit time period is greater than the threshold of the number of people; Calculate the abnormal clustering coefficient of the crowd, the calculation formula is: Among them, YC rq is the abnormal clustering coefficient of the crowd; Geographical location information of the monitored area, including: The regional geographical location information is represented by a safety position deviation coefficient; The logic for obtaining the safety position deviation coefficient is as follows: obtain the floor where the fire source is located, the floor where the monitoring area is located, and the distance between the monitoring area and the safe evacuation area, and obtain the floor deviation value. The calculation formula of the floor deviation value is: Among them, LC pc is the floor deviation value, HY LC is the floor where the fire originated, QY LC is the floor where the monitoring area is located, R is the correction factor, ensuring that the denominator is a positive integer, which is used to adjust the impact of the fire source floor on safety; Obtain the location coordinates of the monitoring area and the safe evacuation area, and calculate the safe evacuation deviation value. The calculation formula for the safe evacuation deviation value is: Among them, CL pc is the safe evacuation deviation value, (x jc ,y jc ) is the location coordinate of the monitoring area, (x cl ,y cl ) are the location coordinates of the safe evacuation area; Calculate the safe position deviation coefficient using the following formula: Among them, PC aq is the safety position deviation coefficient.

2. A high-rise fire intelligent positioning voice alarm method according to claim 1, characterized in that: The concentration of different types of smoke generated by the combustion of different materials, the brightness value of the flame, and the spread speed of the flame are used as the input of the fuzzy Bayesian neural network model, including: Specific gases are identified through specific sensors, monitoring intervals are set, and the average concentrations of different types of smoke in the monitoring intervals are marked as: ND n , where n = 1, 2, 3, ..., N, N is a positive integer, and n is the number of the smoke type; The brightness of the flame is captured by optical sensors, cameras and other equipment, and the average brightness value of the flame in the monitoring interval is marked as: LD; The size and position of each flame frame are captured by the camera to determine the flame spread speed information, which is then expressed by the fire spread accumulation coefficient; The logic for obtaining the fire spread accumulation coefficient is as follows: obtain the initially captured flame image, extract the number of pixels in the flame area in the initial flame image, and mark the number of pixels in the flame area in the initial flame image as: XL cs ; Obtaining each frame of the flame image within the monitoring interval, and extracting the number of pixels belonging to the flame area in each frame of the flame image within the monitoring interval, determining the variation function of the number of pixels in the flame area within the monitoring interval and time according to the time length of the monitoring interval, and marking the variation function of the number of pixels in the flame area within the monitoring interval and time as: XL(t), wherein the variation function of the number of pixels and time is determined by linear or nonlinear fitting; The monitoring interval is marked as: T, and the monitoring interval is divided into the first half and the second half, and the duration of each half is Calculate the fire spread accumulation coefficient using the following formula: Among them, MY jl is the fire spread accumulation coefficient, MY(t)=XL(t)-XL cs MY(t) is the difference function between the number of pixels in the flame area in the monitoring interval and the time and the number of pixels in the flame area in the initial flame image, t a ~t b is the time period in the first half of the monitoring interval that is smaller than the number of pixels in the flame area in the initial flame image, t c ~t d It is the time period in the second half of the monitoring interval where the number of pixels in the flame area is greater than that in the initial flame image.

3. A high-rise fire intelligent positioning voice alarm method according to claim 2, characterized in that: Determine the fire danger score for the fire area, including: The output parameter of the fuzzy Bayesian neural network model is the probability distribution of fire hazard level. The fire hazard level is classified into different levels, including low hazard level, medium hazard level and high hazard level. According to the training results of the fuzzy Bayesian neural network model, the probability distribution of fire hazard level is determined, and the fire hazard score of the fire area is determined; According to the probability distribution of fire hazard level, the fire hazard score is determined by weighted summation calculation. The calculation formula of fire hazard score is: PF hz =α1y d +α2y z +α3y g PF hz Score the fire hazard, y d Low risk level, y z For medium risk level, y g is a high risk level, α1, α2, and α3 are the proportional coefficients of low risk level, medium risk level, and high risk level, respectively, and α1, α2, and α3 are all greater than 0.

4. A high-rise fire intelligent positioning voice alarm method according to claim 3, characterized in that: Quantify the regulation needs of alarms within each monitoring area of ​​the building, including: According to the fire hazard score of the fire, the abnormal crowd aggregation coefficient and the safety position deviation coefficient are weighted and calculated to build an alarm adjustment evaluation model and generate an alarm adjustment evaluation coefficient. The expression of the alarm adjustment evaluation coefficient is: Among them, pg tj is the alarm adjustment evaluation coefficient, β1 and β2 are the proportional coefficients of the abnormal crowd aggregation coefficient and the safe position deviation coefficient, respectively, and β1 and β2 are both greater than 0.

5. A high-rise fire intelligent positioning voice alarm method according to claim 4, characterized in that: Generates alarm adjustment signals in real time, including: Obtain the alarm adjustment evaluation coefficient of each monitoring area in real time, adjust the alarm of the monitoring area according to the alarm adjustment evaluation coefficient of each unit time period of the monitoring area, and compare the alarm adjustment evaluation coefficient of the previous unit time period of the monitoring area with the alarm adjustment evaluation coefficient of the current unit time period; If the alarm adjustment evaluation coefficient of the previous unit time period of the monitoring area is greater than the alarm adjustment evaluation coefficient of the current unit time period, a reduction signal is generated; If the alarm adjustment evaluation coefficient of the previous unit time period of the monitoring area is less than the alarm adjustment evaluation coefficient of the current unit time period, an increase signal is generated.

6. A high-rise fire intelligent positioning voice alarm system, used to implement a high-rise fire intelligent positioning voice alarm method as described in any one of claims 1 to 5, characterized in that: It includes a data acquisition module, a neural network module, an evaluation module and a signal generation module, and the signals between the modules are connected; Data collection module, used to collect real-time information on different types of smoke concentrations, flame brightness values, flame spread speed information, crowd flow information in the monitored area, and regional geographic location information of the monitored area; The neural network module is used to construct a fuzzy Bayesian neural network model to output the probability distribution of the fire hazard level according to the different types of smoke concentration, flame brightness value, and flame spread speed information; An evaluation module is used to construct an alarm adjustment evaluation model based on the crowd flow information and the regional geographical location information of the monitoring area, generate an alarm adjustment evaluation coefficient, and quantify the adjustment requirements of the alarm; The signal generation module is used to adjust the evaluation coefficient of the alarm in different time periods, generate a reduction signal or an increase signal, and realize intelligent adjustment.

Images