Fire-fighting area real-time monitoring system based on Internet of Things

Abstract

The invention relates to the technical field of fire-fighting area management, and discloses a fire-fighting area real-time monitoring system based on the Internet of Things, and the system comprises a data collection module which is used for obtaining the real-time monitoring data of a fire-fighting area; and the central control module is used for establishing an exception recognition model library, selecting a target exception recognition model according to the real-time monitoring data, and generating an adjustment instruction according to the target exception recognition model and the real-time monitoring data. According to the invention, based on an anomaly identification model construction technology, an anomaly identification model is constructed by using historical data according to fire-fighting subareas, and features such as historical data extraction images are analyzed and mined to identify an abnormal event type and an emergency degree; updating the model by using new event information, and adjusting parameters or monitoring strategies according to the frequency and duration of the abnormal event; grouping the real-time monitoring data according to the fire-fighting subareas, and matching the grouped data with the anomaly identification model to judge the state of the fire-fighting subareas; therefore, the emergency degree of the abnormal event can be judged more accurately and a corresponding early warning decision can be made.

Description

Technical Field

[0001] The present invention relates to the technical field of fire area management, and particularly to a real-time monitoring system for fire areas based on the Internet of Things. Background Art

[0002] At present, fire safety is crucial for protecting the lives and property of people. In various building environments, the occurrence of a fire may cause huge damage. Traditional fire monitoring methods have certain limitations, such as it is difficult to achieve real-time, comprehensive, and accurate monitoring, which has promoted the research and development of a real-time monitoring system for fire areas based on Internet of Things technology.

[0003] Fire areas include various fire-fighting facilities, such as fire lanes, safety evacuation channels, fire monitoring rooms, elevators, etc. The normal operation and reasonable use of these facilities are the keys to fire safety. For example, fire lanes must be kept unobstructed, and safety evacuation channels cannot be illegally occupied. Effective management of these facilities and areas requires real-time monitoring to promptly detect and handle abnormal situations. Summary of the Invention

[0004] The object of the present invention is to: conduct real-time monitoring of the entire fire area, including each fire sub-area, different types of fire-related facilities (such as fire lanes, evacuation channels, monitoring rooms, elevators, etc.), and various possible abnormal situations, to ensure that all potential safety hazards in the fire area can be promptly detected.

[0005] To achieve the above object, a real-time monitoring system for fire areas based on the Internet of Things includes:

[0006] A data acquisition module, used to obtain real-time monitoring data of the fire area;

[0007] A central control module, used to establish an abnormal situation recognition model library, select a target abnormal situation recognition model according to the real-time monitoring data, and generate an adjustment instruction according to the target abnormal situation recognition model and the real-time monitoring data;

[0008] Among them, the central control module includes:

[0009] A first control unit: set multiple fire sub-areas based on the fire area, and establish an abnormal situation recognition model for each fire sub-area based on historical data; generate an abnormal situation recognition model library based on the abnormal situation recognition models of all fire sub-areas;

[0010] A second control unit: group the obtained real-time monitoring data, judge the status of the current fire sub-area based on each group of real-time monitoring data, select an abnormal situation recognition model based on the status of the current fire sub-area, and generate a corresponding early warning instruction;

[0011] A third control unit: correct the abnormal situation recognition model library based on the generated early warning instruction.

[0012] In some embodiments of the present invention, the first control unit is further configured to:

[0013] Generate multiple fire compartments based on the building structure of the current fire area obtained, and obtain the historical data of each fire compartment.

[0014] Obtain the historical video data and corresponding event information of the current fire compartment, and extract the image features when the corresponding abnormal event occurs in the historical video data.

[0015] The image features include color features, size features, and position features.

[0016] Identify the current video data based on the color features and size features to generate a set of abnormal objects in the current video data, and the set of abnormal objects includes one or more abnormal objects.

[0017] Generate corresponding abnormal event types based on the abnormal objects in the set of abnormal objects, and combine multiple abnormal event types to generate a set of abnormal events.

[0018] Obtain the development trend of the abnormal objects in the current historical video data based on the position features.

[0019] Set the monitoring strategy for the current abnormal event type based on the development trend of the abnormal objects.

[0020] Calculate the evaluation value of the emergency level of the current abnormal event type based on the residence time of the current abnormal objects, and generate a set of evaluation values of the emergency level P of the set of abnormal events by synthesizing the evaluation values of the emergency levels of each abnormal event type, P = {p 1 , p 2 … p i … p m};

[0021] Among them, p i is the evaluation value of the emergency level of the i-th abnormal event type, and m is the total number of abnormal event types in the set of abnormal events.

[0022] Construct an abnormal recognition model for the current fire compartment based on each event information of the current fire compartment and the corresponding image features.

[0023] Output the set of abnormal events of the current fire compartment and the set of evaluation values of the emergency level P of the set of abnormal events through the abnormal recognition model.

[0024] Generate an abnormal recognition model library M based on the abnormal recognition models of all fire compartments, M = {M 1 , M 2 … M i … M n}, where n is the number of fire compartments.

[0025] M i represents the anomaly recognition model for the i-th fire compartment.

[0026] In some embodiments of the present invention, when obtaining the development trend of the abnormal object in the current historical video data based on the position feature, it includes:

[0027] Obtain the position feature data set D of the abnormal object in the current fire compartment, D = {D 1 , D 2 … D i … D n};

[0028] where D i represents the position feature data of the abnormal object in the current fire compartment at the i-th moment, and n represents the number of position feature data; D i =(x i , y i , z i ), x i represents the data value of the abnormal object in the current fire compartment on the x-axis at the i-th moment; y i represents the data value of the abnormal object in the current fire compartment on the y-axis at the i-th moment; z i represents the data value of the abnormal object in the current fire compartment on the z-axis at the i-th moment;

[0029] Predict the movement trajectory of the current abnormal object based on the calculated displacement vector at the i-th moment, and judge the development trend of the current abnormal object based on the movement trajectory.

[0030] In some embodiments of the present invention, when predicting the movement trajectory of the current abnormal object, it includes:

[0031] Obtain the displacement vectors V k , V 2k , V 3k of the abnormal object from the starting moment to the k-th, 2k-th, and 3k-th moments respectively;

[0032] If any two pairs of offset angles of the abnormal object displacement vectors V k , V 2k , V 3k are less than the angle preset value θ, then judge that the current abnormal object makes the first type of movement;

[0033] Generate an average displacement vector E k , V 2k , V 3k by taking the average of the abnormal object displacement vectors V V ;

[0034] Establish linear prediction equations for the x, y, and z axes based on the average displacement vector E V respectively;

[0035] X = c 1 *x i + d 1 ;

[0036] Y = c 2 *y i + d 2 ;

[0037] Z = c 3 *z i + d 3 ;

[0038] Wherein, c 1 represents the coefficient of the predicted motion value in the x-axis direction, and d 1 represents the constant in the x-axis direction, and c 2 represents the coefficient of the predicted motion value in the y-axis direction, and d 2 represents the constant in the y-axis direction, and c 3 represents the coefficient of the predicted motion value in the z-axis direction, and d 3 represents the constant in the z-axis direction;

[0039] Generate the predicted motion trajectory of the current abnormal object based on the linear prediction equations of the x, y, and z axes;

[0040] If the displacement vector V of the abnormal object k , V 2k , V 3k has an offset angle greater than the preset angle θ for any two pairs, it is determined that the current abnormal object makes a second type of motion.

[0041] In some embodiments of the present invention, when setting the monitoring strategy for the current abnormal event type based on the development trend of the abnormal object, it includes:

[0042] Obtain the development trend of the abnormal object in the current abnormal event and set the first preset time T based on historical data;

[0043] If the development trend of the current abnormal object is the first type of motion, determine the current motion direction based on the predicted motion trajectory of the current abnormal object, and judge whether there is a passage to leave the current fire compartment in combination with the geographical location information of the current fire compartment;

[0044] If there is a passage to leave the current fire compartment in the current motion direction, predict the time T1 when the abnormal object leaves the current fire compartment;

[0045] When T1 < T, execute the first type of monitoring strategy;

[0046] When T1 ≥ T, execute the second type of monitoring strategy;

[0047] If there is no passage to leave the current fire compartment in the current movement direction, the second type of monitoring strategy is executed;

[0048] If the development trend of the current abnormal object is the second type of movement, the second type of monitoring strategy is executed;

[0049] The monitoring frequency of the first type of monitoring strategy is less than that of the second type of monitoring strategy.

[0050] In some embodiments of the present invention, when evaluating the emergency level value of the current abnormal event type, it includes:

[0051] p = f1 * t1 / T;

[0052] Wherein, t1 is the real-time residence time of the current abnormal object in the fire compartment, f1 is the first time coefficient, and p is the emergency level evaluation value of the current abnormal event type.

[0053] In some embodiments of the present invention, the second control unit is further configured to:

[0054] Group the real-time monitoring data obtained from the data acquisition module according to the fire compartment to generate grouped real-time monitoring data;

[0055] Input the grouped real-time monitoring data into the abnormal recognition model to output the abnormal event set of the current fire compartment and the set P of the emergency level evaluation values of the abnormal event set;

[0056] Formulating an emergency level preset value based on historical emergency level evaluation value data, including: the first preset value a1 and the second preset value a2;

[0057] Compare the set P of the emergency level evaluation values of the abnormal event set, P = {p 1 , p 2 … p i … p m} with the emergency level preset value;

[0058] Wherein, p i is the emergency level evaluation value of the i-th abnormal event, and m is the total number of abnormal events in the abnormal event set;

[0059] If p i < a1, then monitor and predict the change of the evaluation value in real time;

[0060] If a1 ≤ p i ≤ a2, then correct the emergency level evaluation value of the current abnormal event type based on the obtained infrared data to generate the corrected value PX of the emergency level evaluation value of the current abnormal event type i Compare with the emergency level preset value again;

[0061] If a2 < p i, a corresponding warning instruction is generated.

[0062] In some embodiments of the present invention, when correcting the evaluation value of the urgency of the current abnormal event type, it includes:

[0063] Based on the infrared data, select the abnormal event types in the abnormal event set to determine the abnormal event type in the current fire area;

[0064] Combined with the current abnormal event type and historical data,

[0065] Obtain the warning times b1 corresponding to the current abnormal event type in the historical data and the total times B of the current abnormal event type;

[0066] Generate a correction value PX for the evaluation value of the urgency of the current abnormal event type i ,

[0067] where E is the correction value coefficient of the evaluation value of the urgency of the current abnormal event type.

[0068] In some embodiments of the present invention, the third control unit further includes:

[0069] Add the processed event information to the abnormal recognition model library, obtain the occurrence frequency and duration of the abnormal event, and execute the first monitoring strategy for the abnormal event whose occurrence frequency in the historical data exceeds the preset value.

[0070] Compared with the prior art, the beneficial effects of a real-time monitoring system for a fire area based on the Internet of Things provided by the embodiments of the present invention are as follows:

[0071] The abnormal recognition model in the central control module is constructed based on the historical data of the fire partition, considering the specific conditions of each fire partition and the development trend of abnormal objects in different abnormal events, so as to more accurately evaluate the state of the fire area.

[0072] The second control unit can quickly generate a warning instruction according to the real-time monitoring data and the abnormal recognition model; once an abnormal event is detected, the system can notify relevant personnel in a short time.

[0073] Set different monitoring strategies based on the development trend of abnormal objects. The monitoring strategies flexibly adjusted according to the actual situation can reasonably allocate monitoring resources, ensure that the abnormal events most likely to cause serious consequences can be focused on in an emergency, and improve the efficiency of emergency response.

[0074] The method of dynamically adjusting the monitoring frequency according to the urgency level avoids the indiscriminate and high-frequency monitoring of all monitoring areas and events, saves monitoring resources, and at the same time ensures the key attention to high-risk areas and events.

[0075] The third control unit adds the processed event information to the abnormal recognition model library, and updates and optimizes the model according to the occurrence frequency and duration of abnormal events; more reasonably allocates monitoring resources and improves the effectiveness of monitoring. Description of the Drawings

[0076] Figure 1 is a flowchart of a real-time fire area monitoring system based on the Internet of Things provided by an embodiment of the present invention. Detailed Embodiments

[0077] The following combines the drawings and embodiments to further describe in detail the specific embodiments of the present invention. The following embodiments are used to illustrate the present invention, but are not used to limit the scope of the present invention.

[0078] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation of the present invention.

[0079] The terms "first" and "second" are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present invention, unless otherwise stated, the meaning of "plurality" is two or more.

[0080] In the description of the present invention, it should be noted that unless otherwise clearly defined and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense. For example, it may be a fixed connection, a detachable connection, or an integral connection; it may be a mechanical connection or an electrical connection; it may be directly connected or indirectly connected through an intermediate medium, and it may be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0081] Embodiment 1:

[0082] A real-time fire area monitoring system based on the Internet of Things provided by an embodiment of the present invention, as Figure 1As shown in the figure, it includes:

[0083] A data acquisition module, used to obtain real-time monitoring data of the fire area;

[0084] A central control module, used to establish an abnormal recognition model library, select a target abnormal recognition model according to the real-time monitoring data, and generate an adjustment instruction according to the target abnormal recognition model and the real-time monitoring data;

[0085] Among them, the central control module includes:

[0086] The first control unit: set multiple fire compartments based on the fire area, and establish an abnormal recognition model for each fire compartment based on historical data; generate an abnormal recognition model library based on the abnormal recognition models of all fire compartments;

[0087] The second control unit: group the obtained real-time monitoring data, judge the status of the current fire compartment based on each group of real-time monitoring data, select an abnormal recognition model based on the status of the current fire compartment, and generate a corresponding warning instruction;

[0088] The third control unit: correct the abnormal recognition model library based on the generated warning instruction.

[0089] Embodiment 2:

[0090] The first control unit is also used for:

[0091] Generate multiple fire compartments based on the building structure of the current fire area obtained, and obtain the historical data of each fire compartment,

[0092] Obtain the historical video data and corresponding event information of the current fire compartment, and extract the image features when the corresponding abnormal event occurs in the historical video data;

[0093] The image features include color features, size features, and position features;

[0094] Based on the color features and size features, identify the current video data to generate a set of abnormal objects in the current video data, and the set of abnormal objects includes one or more abnormal objects;

[0095] Generate corresponding abnormal event types based on the abnormal objects in the set of abnormal objects, and generate an abnormal event set by combining multiple abnormal event types;

[0096] Obtain the development trend of the abnormal objects in the current historical video data based on the position features;

[0097] Set the monitoring strategy for the current abnormal event type based on the development trend of the abnormal objects;

[0098] Calculate the emergency level evaluation value of the current abnormal event type based on the residence time of the current abnormal object, and generate the emergency level evaluation value set P of the abnormal event set by integrating the emergency level evaluation values of each abnormal event type. P = {p 1 , p 2 … p i … p m};

[0099] Among them, p i is the emergency level evaluation value of the i-th abnormal event type, and m is the total number of abnormal event types in the abnormal event set;

[0100] Construct an abnormal recognition model for the current fire compartment based on each event information in the current fire compartment and the image features corresponding to the event information;

[0101] Output the abnormal event set of the current fire compartment and the emergency level evaluation value set P of the abnormal event set through the abnormal recognition model;

[0102] Generate an abnormal recognition model library M according to the abnormal recognition models of all fire compartments. M = {M 1 , M 2 … M i … M n};

[0103] M i represents the abnormal recognition model of the i-th fire compartment.

[0104] Example 3:

[0105] When obtaining the development trend of abnormal objects in the current historical video data based on location features, it includes:

[0106] Obtain the location feature data set D of abnormal objects in the current fire compartment. D = {D 1 , D 2 … D i … D n};

[0107] Among them, D i represents the location feature data of abnormal objects in the current fire compartment at the i-th moment, and n represents the number of location feature data; D i = (x i , y i , z i ), x i represents the data value of the abnormal object in the current fire compartment on the x-axis at the i-th moment; y i represents the data value of the abnormal object in the current fire compartment on the y-axis at the i-th moment; z i represents the data value of the abnormal object in the current fire compartment on the z-axis at the i-th moment;

[0108] Predict the movement trajectory of the current abnormal object based on the calculated displacement vector at the i-th moment, and judge the development trend of the current abnormal object based on the movement trajectory.

[0109] In this embodiment, first obtain the position feature dataset D of the abnormal object in the current fire compartment. According to these position feature data, the displacement vector at the i-th moment can be calculated. By analyzing and processing the displacement vector, the movement trajectory of the current abnormal object can be predicted. For example, the displacement vector can be calculated using the difference in position data at adjacent moments, and then the future movement path of the abnormal object can be predicted based on the pattern of a series of displacement vectors.

[0110] Judge the development trend of the current abnormal object based on the predicted movement trajectory. If the movement trajectory shows that the abnormal object moves stably in a specific direction, it may indicate a certain tendency of diffusion or spread of the abnormal object; if the movement trajectory is relatively chaotic, it may indicate that the development of the abnormal object is uncertain or in a state of irregular activity. This judgment of the development trend of the abnormal object helps to formulate corresponding countermeasures. The development trend of all current abnormal objects includes the movement type, movement trajectory, and residence time of the current abnormal object.

[0111] Embodiment 4:

[0112] When predicting the movement trajectory of the current abnormal object, it includes:

[0113] Obtain the displacement vectors V of the abnormal object at the starting moment to the k-th, 2k-th, and 3k-th moments respectively k , V 2k , V 3k ;

[0114] If any two pairs of offset angles of the abnormal object displacement vectors V k , V 2k , V 3k are less than the angle preset value θ, then judge that the current abnormal object makes the first type of movement;

[0115] Based on the abnormal object displacement vectors V k , V 2k , V 3k Take the average value to generate the average displacement vector E V ;

[0116] Based on the average displacement vector E V Establish linear prediction equations for the x, y, and z axes respectively;

[0117] X = c 1 *x i +d 1 ;

[0118] Y = c 2 *yi +d 2 ;

[0119] Z = c 3 *z i +d 3 ;

[0120] Wherein, c 1 represents the coefficient of the predicted motion value in the x-axis direction, and d 1 represents the constant in the x-axis direction, and c 2 represents the coefficient of the predicted motion value in the y-axis direction, and d 2 represents the constant in the y-axis direction, and c 3 represents the coefficient of the predicted motion value in the z-axis direction, and d 3 represents the constant in the z-axis direction;

[0121] Generate the predicted motion trajectory of the current abnormal object based on the linear prediction equations of the x, y, and z axes;

[0122] If any two pairs of offset angles of the abnormal object displacement vector V k , V 2k , V 3k are greater than the angle preset value θ, it is determined that the current abnormal object makes a second type of motion.

[0123] In this embodiment, the abnormal object displacement vectors V k , V 2k , V 3k at the starting time to the k, 2k, and 3k times are respectively obtained. By comparing the offset angles between these displacement vectors in pairs with the angle preset value θ, it can be determined whether the motion of the abnormal object is regular.

[0124] If any two pairs of offset angles of the abnormal object displacement vector V k , V 2k , V 3k are less than the angle preset value θ, this indicates that the motion direction of the abnormal object is relatively stable at these times, so it is determined that the current abnormal object makes a first type of motion. On the contrary, if any two pairs of offset angles are greater than the angle preset value θ, it is determined that the current abnormal object makes a second type of motion.

[0125] When it is determined to be the first type of motion, based on the abnormal object displacement vectors V k , V 2k , V 3k take the average value to generate the average displacement vector E V . This average displacement vector can reflect the overall motion trend of the abnormal object.

[0126] Based on the average displacement vector E VLinear prediction equations for the x, y, and z axes are established respectively, where the coefficients c1, c2, c3 and the constants d1, d2, d3 are determined according to the specific movement of the abnormal object.

[0127] Finally, based on the linear prediction equations for the x, y, and z axes, the predicted movement trajectory of the current abnormal object is generated. These equations can predict the position of the abnormal object in space at a future time based on the values at the current moment, thereby obtaining its movement trajectory.

[0128] Embodiment 5:

[0129] When setting the monitoring strategy for the current abnormal event type based on the development trend of the abnormal object, it includes:

[0130] Obtain the development trend of the abnormal object in the current abnormal event and set the first preset time T based on historical data;

[0131] If the development trend of the current abnormal object is the first type of movement, determine the current movement direction based on the predicted movement trajectory of the current abnormal object, and combine the geographical location information of the current fire compartment to determine whether there is a passage to leave the current fire compartment in the current movement direction;

[0132] If there is a passage to leave the current fire compartment in the current movement direction, predict the time T1 when the abnormal object leaves the current fire compartment;

[0133] When T1 < T, execute the first type of monitoring strategy;

[0134] When T1 ≥ T, execute the second type of monitoring strategy;

[0135] If there is no passage to leave the current fire compartment in the current movement direction, execute the second type of monitoring strategy;

[0136] If the development trend of the current abnormal object is the second type of movement, execute the second type of monitoring strategy;

[0137] The monitoring frequency of the first type of monitoring strategy is less than that of the second type of monitoring strategy.

[0138] In this embodiment, first, the system will obtain the development trend of the abnormal object in the current abnormal event and set the first preset time T based on historical data. This preset time T is determined according to the development of past similar abnormal events and can be used as a reference standard for evaluating the urgency of the current abnormal event.

[0139] If the development trend of the current abnormal object is the first type of movement, the system will determine the current movement direction based on the predicted movement trajectory of the current abnormal object. This step requires combining the geographical location information of the current fire compartment to determine whether there is a passage to leave the current fire compartment in the current movement direction.

[0140] If there is a passage leading out of the current fire compartment, the system will predict the time T1 when the abnormal object leaves the current fire compartment. When T1 < T, the first type of monitoring strategy is executed. This means that if the abnormal object is expected to leave the current fire compartment in a relatively short time, a relatively loose monitoring strategy can be adopted because the abnormal object may not pose a direct threat to the current compartment.

[0141] In other cases, the second type of monitoring strategy is executed. These cases include that there is no passage leading out of the current fire compartment in the current movement direction, there is a passage leading out of the current fire compartment in the current movement direction but T1 ≥ T, and the development trend of the current abnormal object is the second type of movement. In these cases, the abnormal object may pose a threat to the current fire compartment, so a more stringent monitoring strategy is required.

[0142] The first type of movement is regular movement, and its trajectory can be predicted. The second type of movement is irregular movement and its development trend cannot be predicted, so the monitoring frequency cannot be reduced.

[0143] The monitoring frequency of the first type of monitoring strategy is less than that of the second type of monitoring strategy. This is because the first type of monitoring strategy is applicable to the situation where the abnormal object may quickly leave the current compartment, so it is not necessary to monitor too frequently. While the second type of monitoring strategy is applicable to the situation where the abnormal object may pose a threat to the current compartment, and closer monitoring is required to ensure timely detection and handling of any potential hazards.

[0144] Example 6:

[0145] When evaluating the emergency level value of the current abnormal event type, it includes:

[0146] p = f1 * t1 / T;

[0147] Wherein, t1 is the real-time residence time of the current abnormal object in the fire compartment, f1 is the first time coefficient, and p is the emergency level evaluation value of the current abnormal event type.

[0148] In this embodiment, in the fire compartment, the real-time residence time t1 of the current abnormal object is an important factor in measuring the emergency level of the abnormal event. The longer the residence time, the greater the potential harm that the abnormal event may cause to the area. Abnormal events such as illegally occupying the fire lane; illegally occupying the safety evacuation passage; no one on duty in the fire monitoring room; illegally entering the elevator.

[0149] The first time coefficient f1 is a parameter used to adjust the emergency level evaluation value. Its value may be based on the characteristics of the fire compartment.

[0150] Example 7:

[0151] The second control unit is further used for:

[0152] Group the real-time monitoring data obtained from the data acquisition module according to the fire compartments to generate the grouped real-time monitoring data;

[0153] Input the grouped real-time monitoring data into the anomaly recognition model to output the set of anomaly events in the current fire compartment and the set P of evaluation values of the emergency levels of the set of anomaly events;

[0154] Set the preset emergency level values based on the historical evaluation data of emergency levels, including: the first preset value a1 and the second preset value a2;

[0155] Compare the set P of evaluation values of the emergency levels of the set of anomaly events, P = {p 1 , p 2 … p i … p m} with the preset emergency level values;

[0156] If p i < a1, then monitor and predict the change of the evaluation value in real time;

[0157] If a1 ≤ p i ≤ a2, then correct the evaluation value of the emergency level of the current anomaly event type based on the obtained infrared data to generate the corrected value PX of the evaluation value of the emergency level of the current anomaly event type i Compare with the preset emergency level values again;

[0158] If a2 < p i , then generate the corresponding warning instruction.

[0159] In this embodiment, grouping the real-time monitoring data obtained from the data acquisition module according to the fire compartments to generate the grouped real-time monitoring data helps to manage the monitoring data in an organized manner. Different fire compartments may have different characteristics and risk factors. After grouping, independent analysis and processing can be carried out for each compartment, improving the accuracy and efficiency of anomaly recognition.

[0160] Input the grouped real-time monitoring data into the anomaly recognition model to output the set of anomaly events in the current fire compartment and the set P of evaluation values of the emergency levels of the set of anomaly events. The anomaly recognition model can identify possible obstacles that hinder the fire compartment based on various features in the monitoring data (such as pedestrians, vehicles, etc.) and quantitatively evaluate the emergency levels of these anomaly events.

[0161] When \(a1\leq pi\leq a2\), the evaluation value of the emergency level of the current abnormal event type is corrected based on the acquired infrared data, and a corrected value \(PXi\) of the evaluation value of the emergency level of the current abnormal event type is generated and compared with the preset value of the emergency level again. This indicates that the abnormal event is in the medium emergency level range. The infrared data may provide more detailed information to more accurately evaluate the emergency level, and making a comparison again through the corrected value can lead to a more accurate decision.

[0162] Embodiment 8:

[0163] When correcting the evaluation value of the emergency level of the current abnormal event type, it includes:

[0164] Select the abnormal event type in the abnormal event set based on the infrared data to determine the abnormal event type in the current fire area;

[0165] Combined with the current abnormal event type and historical data,

[0166] Obtain the number of early warning times \(b1\) corresponding to the current abnormal event type in the historical data and the total number \(B\) of the current abnormal event type;

[0167] Generate a corrected value \(PX\) of the evaluation value of the emergency level of the current abnormal event type i ,

[0168] where \(E\) is the correction coefficient of the evaluation value of the emergency level of the current abnormal event type.

[0169] In this embodiment, the infrared data is of great significance in determining the abnormal event type in the current fire area. Infrared technology can detect the thermal radiation of objects. In a fire protection scenario, different abnormal events (such as fires, overheating of electrical equipment, etc.) will generate different thermal characteristics.

[0170] By analyzing the infrared data, different abnormal event types in the abnormal event set can be distinguished, so as to accurately determine the abnormal event type occurring in the current fire area. For example, a fire will show a large area of high-temperature region, while overheating of electrical equipment may be a local high-temperature point.

[0171] Obtaining the number of early warning times \(b1\) corresponding to the current abnormal event type in the historical data and the total number \(B\) of the current abnormal event type is to evaluate the historical occurrence situation of this abnormal event type.

[0172] The number of early warning times \(b1\) reflects the frequency of issuing early warnings for this abnormal event type in history, while the total number \(B\) gives the overall occurrence situation of this abnormal event type. These two data can help us analyze the severity and occurrence regularity of this abnormal event type.

[0173] For example, if the ratio of the number of warning times b1 of an abnormal event type to the total number B is relatively high, it indicates that this event was considered relatively serious in the past and often required warnings.

[0174] Generate the correction value PX for the emergency level evaluation value of the current abnormal event type i , where E is the correction coefficient. The role of E is to adjust the magnitude of the correction value calculated based on the number of warning times and the total number.

[0175] The value of E may be determined according to the specific requirements of the fire protection system and the actual situation. If E is larger, it indicates that more importance is attached to the influence of historical data on the current emergency level evaluation value; if E is smaller, the influence of historical data is relatively small.

[0176] The correction value PX i may be calculated based on a certain functional relationship between the number of warning times b1 and the total number B and multiplied by the correction coefficient E. For example, PX i = E * (b1 / B), so that the emergency level evaluation value of the current abnormal event type can be reasonably corrected according to historical data, making the evaluation of the emergency level more accurate and in line with the actual situation.

[0177] Embodiment 9:

[0178] The third control unit further includes:

[0179] Add the processed event information to the abnormal recognition model library, obtain the occurrence frequency of abnormal events and the duration of abnormal events, and execute the first monitoring strategy for abnormal events whose occurrence frequency in historical data exceeds the preset value.

[0180] In this embodiment, adding the processed event information to the abnormal recognition model library can continuously expand the data volume of the model library. This is very important for the abnormal recognition model, just like providing more learning materials for a student. As the data increases, the model can better learn various characteristic patterns of abnormal events, thereby improving the recognition accuracy and generalization ability.

[0181] Obtaining the occurrence frequency of abnormal events and the duration of abnormal events can be used as an important basis for evaluating the risk level of abnormal events. For example, if the occurrence frequency of an abnormal event in historical data is very high, the monitoring frequency for common abnormalities can be reduced, such as students passing through the campus fire access road.

[0182] Finally, it should be noted that: Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and its equivalent technologies, the present invention is also intended to include these changes and modifications.

[0183] The above are only the preferred embodiments of the present invention and are not intended to limit the protection scope of the present invention.

Claims

1. A real-time monitoring system for fire protection areas based on the Internet of Things, characterized in that: include: Data acquisition module, used to obtain real-time monitoring data of the fire protection area; The central control module is used to establish an abnormality recognition model library, select a target abnormality recognition model according to real-time monitoring data, and generate adjustment instructions according to the target abnormality recognition model and real-time monitoring data; Wherein, the central control module includes: The first control unit: sets multiple fire zones based on the fire area, and establishes an abnormality recognition model for each fire zone based on historical data; generates an abnormality recognition model library based on the abnormality recognition models of all fire zones; The second control unit: groups the acquired real-time monitoring data, determines the status of the current fire zone based on each group of real-time monitoring data, selects an abnormality recognition model based on the status of the current fire zone, and generates a corresponding warning instruction; The third control unit: amends the abnormality recognition model library based on the generated warning instruction.

2. The real-time monitoring system for fire protection areas based on the Internet of Things as claimed in claim 1, characterized in that: The first control unit is also used for: Generate multiple fire zones based on the building structure of the current fire zone, and obtain historical data of each fire zone. Obtain the historical video data and corresponding event information of the current fire zone, and extract the image features of the corresponding abnormal events in the historical video data; Image features include color features, size features, and position features; Identify the current video data based on the color feature and the size feature to generate an abnormal object set in the current video data, where the abnormal object set includes one or more abnormal objects; Generate corresponding abnormal event types based on abnormal objects in the abnormal object set, and generate abnormal event sets by combining multiple abnormal event types; Obtain the development trend of abnormal objects in current historical video data based on location features; Set monitoring strategies for current abnormal event types based on the development trend of abnormal objects; The urgency evaluation value of the current abnormal event type is calculated based on the residence time of the current abnormal object, and the urgency evaluation value set P of the abnormal event set is generated by combining the urgency evaluation values ​​of each abnormal event type, where P = {p1, p2…p i …p m }; Among them, p i is the urgency evaluation value of the ith abnormal event type, and m is the total number of abnormal event types in the abnormal event set; An abnormality recognition model for the current fire zone is constructed based on each event information of the current fire zone and the image features corresponding to the event information; Output the abnormal event set of the current fire zone and the emergency evaluation value set P of the abnormal event set through the abnormal recognition model; Generate anomaly recognition model library M based on anomaly recognition models of all fire zones, M = {M1, M2…M i …M n }; M i represents the anomaly recognition model of the i-th fire zone, and n is the number of fire zones.

3. The real-time monitoring system for fire protection areas based on the Internet of Things as claimed in claim 2, characterized in that: The method of obtaining the development trend of abnormal objects in the current historical video data based on the location feature includes: Obtain the location feature dataset D of the abnormal objects in the current fire zone, D = {D1, D2…D i …D n }; Among them, D i represents the location feature data of the abnormal object in the current fire zone at the i-th moment, and n represents the number of location feature data; D i =(x i ,y i ,z i ), x i represents the data value of the abnormal object in the current fire zone at the i-th moment on the x-axis; y i represents the data value of the abnormal object on the y-axis in the current fire zone at the i-th moment; z i Represents the data value of the abnormal object on the z-axis in the current fire zone at the i-th moment; Based on the calculated displacement vector at the i-th moment, the motion trajectory of the current abnormal object is predicted, and the development trend of the current abnormal object is determined based on the motion trajectory.

4. The real-time monitoring system for fire protection areas based on the Internet of Things as claimed in claim 3 is characterized in that: The predicting of the movement trajectory of the current abnormal object includes: Get the abnormal object displacement vector V from the starting time to time k, 2k and 3k respectively k , V 2k , V 3k ; If the abnormal object displacement vector V k , V 2k , V 3k If any two pairs of offset angles are less than the preset angle value θ, it is determined that the current abnormal object is performing the first type of motion; Based on the abnormal object displacement vector V k , V 2k , V 3k Take the average value to generate the average displacement vector E V ; Based on the average displacement vector E V Establish linear prediction equations for x, y, and z axes respectively; X=c1*x i +d1; Y=c2*y i +d2; Z=c3*z i +d3; Wherein, c1 represents the coefficient of the predicted motion value in the x-axis direction, d1 represents the constant in the x-axis direction, c2 represents the coefficient of the predicted motion value in the y-axis direction, d2 represents the constant in the y-axis direction, c3 represents the coefficient of the predicted motion value in the z-axis direction, and d3 represents the constant in the z-axis direction; Generate the predicted motion trajectory of the current anomaly based on the linear prediction equations of the x, y, and z axes; If the abnormal object displacement vector V k , V 2k , V 3k If the offset angle between any two teams is greater than the preset angle value θ, it is determined that the current abnormal object is performing the second type of motion.

5. The real-time monitoring system for fire protection areas based on the Internet of Things as claimed in claim 4, characterized in that: The monitoring strategy for the current abnormal event type is set based on the development trend of the abnormal object, including: Obtain the development trend of abnormal objects in the current abnormal event and set a first preset time T based on historical data; If the current abnormal object performs the first type of motion, the current motion direction is determined based on the predicted motion trajectory of the current abnormal object, and combined with the geographical location information of the current fire zone, it is determined whether there is a passage to leave the current fire zone in the current motion direction; If there is a passage out of the current fire compartment in the current movement direction, predict the time T1 when the abnormal object leaves the current fire compartment; When T1 < T, execute the first type of monitoring strategy; When T1 ≥ T, execute the second type of monitoring strategy; If there is no passage out of the current fire compartment in the current movement direction, execute the second type of monitoring strategy; If the current abnormal object makes the second type of movement, execute the second type of monitoring strategy; The monitoring frequency of the first type of monitoring strategy is less than that of the second type of monitoring strategy.

6. The real-time monitoring system for fire protection areas based on the Internet of Things as claimed in claim 5, characterized in that: When evaluating the emergency level of the current abnormal event type, it includes: p = f1 * t1 / T; Where, t1 is the real-time residence time of the current abnormal object in the fire compartment, f1 is the first time coefficient, and p is the emergency level evaluation value of the current abnormal event type.

7. The real-time monitoring system for fire protection areas based on the Internet of Things as claimed in claim 6, characterized in that: The second control unit is also used for: Group the real-time monitoring data obtained from the data acquisition module according to the fire compartment to generate grouped real-time monitoring data; Input the grouped real-time monitoring data into the abnormal recognition model, and output the monitoring strategy of the current fire compartment, the abnormal event set, and the set P of the emergency level evaluation values of the abnormal event set; Formulate the emergency level preset value based on the historical emergency level evaluation value data, including: the first preset value a1 and the second preset value a2; Compare the emergency evaluation value set P of the abnormal event set, P = {p1, p2…p i …p m } and the urgency preset value; Among them, p i is the urgency evaluation value of the ith abnormal event, and m is the total number of abnormal events in the abnormal event set; If p i <a1, then the change of the real-time monitoring prediction evaluation value is monitored; If a1≤p i ≤a2, the emergency evaluation value of the current abnormal event type is corrected based on the acquired infrared data to generate a correction value PX of the emergency evaluation value of the current abnormal event type i Compare again with the preset value of urgency; If a2 <p i , then generate the corresponding warning instruction.

8. The real-time monitoring system for fire protection areas based on the Internet of Things as claimed in claim 7, characterized in that: When correcting the emergency level evaluation value of the current abnormal event type, it includes: Select the abnormal event type in the abnormal event set based on the infrared data to determine the abnormal event type of the current fire area; Obtain the early warning times b1 corresponding to the current abnormal event type and the total times B of all abnormal event types in the historical data; Generate the current abnormal event type emergency evaluation value correction value PX i , Where, E is the correction coefficient of the emergency level evaluation value of the current abnormal event type.

9. The real-time monitoring system for fire protection areas based on the Internet of Things as claimed in claim 8, characterized in that: The third control unit also includes: Add the processed event information to the abnormal recognition model library, obtain the occurrence frequency of the abnormal event and the duration of the abnormal event, and execute the first monitoring strategy for the abnormal events whose occurrence frequency exceeds the preset value in the historical data.

Images