A method and system for early warning of emergencies based on emergency portraits

Abstract

The present invention belongs to the technical field of emergency management. A method and system for early warning of emergencies based on emergency portraits are provided. The method comprises: collecting raw data corresponding to each type of emergency, extracting first element features and fusion calculation of the raw data using a first fine granularity to obtain a first portrait feature, and constructing an emergency portrait model based on the first target portrait feature; extracting second element features and fusion calculation of real-time multi-source data to obtain a number of second portrait features; inputting all second portrait features into the emergency portrait model to generate an emergency evolution trend corresponding to each type of emergency; when the emergency evolution trend meets the early warning condition, outputting emergency early warning information corresponding to the emergency type. The present invention accurately predicts whether an emergency is about to occur based on the emergency portrait, and improves the efficiency and effect of handling emergencies through early warning.

Description

Technical Field

[0001] The present invention relates to the technical field of emergency management, and in particular to an emergency early warning method and system based on emergency profiling. Background Art

[0002] Information on emergencies is crucial to emergency management, which includes multiple links from collection, processing, transmission to release.

[0003] The following existing patent technologies were found through searching: CN115187148A (Emergency Situation Analysis Method, System, Device and Readable Storage Medium), which discloses an emergency situation analysis method, which includes: extracting scenario keywords from the reported information of the emergency through intelligent analysis service, and determining the event scenario and consequence scenario according to the scenario keywords, wherein the scenario keywords include key parameters; generating a scenario portrait according to the scenario keywords and the information of the involved subject associated with the scenario keywords; inputting the key parameters into the simulation model service to determine the secondary derivative risk and the secondary derivative consequence; The system can also be used to determine the risk level corresponding to the secondary risk; evaluate the emergency resource demand according to the event scenario and the consequence scenario, and analyze the resource gap list according to the resource reserve situation; obtain the on-site panoramic image of the emergency based on the drone panoramic mapping service, map the consequence points, risk points and resource points on the on-site panoramic image, and generate a thematic layer; determine the public opinion index of the emergency; and generate an emergency situation analysis report based on the intelligent analysis results, the scenario portrait, the secondary derivative risks, the risk level corresponding to the secondary derivative risks, the resource reserve situation, the resource gap list, the public opinion index and the thematic layer.

[0004] The patent document uses the portrait technology to analyze the development trend of emergencies, but lacks a more detailed application of the portrait technology, resulting in insufficient personalization and accuracy in the analysis of emergencies. The present invention aims to apply the portrait technology to the early warning of emergencies, and further considers many actual situations to improve the accuracy of the early warning of emergencies. Summary of the invention

[0005] In this regard, the present invention provides an emergency warning method, system, electronic device, computer storage medium and computer program product based on emergency portrait to solve the above technical problems.

[0006] The present invention discloses an emergency warning method based on emergency portrait, which comprises the following steps: determining a number of emergency types and corresponding first fine-grainedness, collecting raw data corresponding to each of the emergency types, and using the first fine-grainedness to extract first element features from the raw data; performing fusion calculation on the extracted first element features to obtain first portrait features, and constructing an emergency portrait model based on the first target portrait features; acquiring real-time multi-source data corresponding to any of the emergency types through multi-source channels, performing second element feature extraction on the real-time multi-source data, and performing fusion calculation on the extracted second element features to obtain a number of second portrait features; inputting all of the second portrait features into the emergency portrait model to generate an emergency evolution trend corresponding to each of the emergency types; and outputting emergency warning information corresponding to the emergency type when the emergency evolution trend meets the warning condition.

[0007] In some embodiments, the determining of several emergency event types and the corresponding first fine-grainedness includes: receiving several emergency event types set manually, evaluating the emergency degree of each emergency event type according to preset evaluation rules, and obtaining the emergency degree corresponding to each emergency event type; and determining the first fine-grainedness corresponding to each emergency event type according to the emergency degree.

[0008] In some embodiments, determining the first fine granularity corresponding to each of the emergency event types according to the emergency degree is specifically: ; In the above formula, For the first fine granularity, is a constant value, used to represent the regular value of the first fine-grainedness; The emergency degree corresponding to any of the emergency event types; is the adjustment parameter and is the preset value.

[0009] In some embodiments, the second element feature extraction of the real-time multi-source data includes: obtaining historical data of emergencies corresponding to the type of emergencies in the monitoring area, and deriving an event probability distribution function corresponding to the type of emergencies in the monitoring area based on statistics of the historical data of emergencies, wherein the event probability distribution function is used to characterize the time distribution of the probability of occurrence of events of a specific type of emergencies in the monitoring area; determining the probability of occurrence of an event corresponding to the current time period based on the event probability distribution function, and determining a second fine granularity based on the probability of occurrence of the event; and extracting the second element feature of the real-time multi-source data according to the second fine granularity.

[0010] In some embodiments, the second fine granularity is determined according to the probability of occurrence of the event: ; In the above formula, For the second fine granularity, is a constant value, used to represent the regular value of the second finer granularity; The probability of occurrence of the emergency event type in the monitoring area in the current time period; is the adjustment parameter and is the preset value.

[0011] In some embodiments, all of the second portrait features are input into the emergency event portrait model to generate an emergency event evolution trend corresponding to each of the emergency event types, including: grouping all of the second portrait features in the order of collection and inputting them into the emergency event portrait model, the emergency event portrait model performing predictive analysis on each group of the second portrait features to obtain the corresponding event prediction probability of occurrence; performing curve fitting on all of the event prediction probabilities in accordance with the above-mentioned order of collection to obtain an event prediction probability curve, which is the emergency event evolution trend; wherein the number of groups is determined according to the second fine-grainedness, and the number of groups is positively correlated with the second fine-grainedness.

[0012] The present invention also discloses an emergency warning system based on emergency portrait, the system comprising a first extraction module, a modeling module, a second extraction module, a prediction module, and an early warning module; the first extraction module is used to determine a number of emergency types and corresponding first fine-grainedness, collect original data corresponding to each of the emergency types, and use the first fine-grainedness to extract first element features from the original data; the modeling module is used to perform fusion calculation on the extracted first element features to obtain first portrait features, and construct an emergency portrait model based on the first target portrait features; the second extraction module is used to obtain real-time multi-source data corresponding to any of the emergency types through multi-source channels, perform second element feature extraction on the real-time multi-source data, perform fusion calculation on the extracted second element features to obtain a number of second portrait features; the prediction module is used to input all the second portrait features into the emergency portrait model to generate an emergency evolution trend corresponding to each of the emergency types; the early warning module is used to output emergency warning information corresponding to the emergency type when the emergency evolution trend meets the early warning condition.

[0013] The present invention also discloses an electronic device, comprising: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the processor executes the computer program to implement the method described in any of the preceding items.

[0014] The present invention also discloses a computer storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement any of the above methods.

[0015] The present invention also discloses a computer program product. When the computer program product is run on a terminal, the terminal implements any of the above methods when executing the computer program product.

[0016] The beneficial effects of the present invention are as follows: the present invention analyzes multi-source data collected in real time based on the constructed emergency portrait model, and obtains the emergency evolution trend of the corresponding emergency type. According to the emergency evolution trend, it can be predicted in advance whether the corresponding type of emergency is about to occur, and relevant personnel or equipment can be dealt with in time through early warning, thereby improving the efficiency and effect of handling emergencies. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 It is a flow chart of an emergency warning method based on emergency profiling disclosed in an embodiment of the present invention.

[0019] Figure 2 It is a schematic diagram of a flow chart of generating an emergency event evolution trend disclosed in an embodiment of the present invention.

[0020] Figure 3 It is a structural schematic diagram of an emergency warning system based on emergency portrait disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0021] The following is a description of the implementation of the present application by specific specific embodiments. People familiar with the technology can easily understand other advantages and effects of the present application from the contents disclosed in this specification. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0022] In addition, the technical features involved in the different embodiments of the present application described below can be combined with each other as long as they do not conflict with each other.

[0023] like Figure 1As shown, an embodiment of the present invention discloses an emergency warning method based on emergency portrait, and the method includes the following steps: determining several emergency types and corresponding first fine-grainedness, collecting raw data corresponding to each of the emergency types, and using the first fine-grainedness to extract first element features from the raw data; fusing and calculating the extracted first element features to obtain first portrait features, and constructing an emergency portrait model based on the first target portrait features; acquiring real-time multi-source data corresponding to any of the emergency types through multi-source channels, extracting second element features from the real-time multi-source data, and fusing and calculating the extracted second element features to obtain several second portrait features; inputting all of the second portrait features into the emergency portrait model to generate an emergency evolution trend corresponding to each of the emergency types; and when the emergency evolution trend meets the warning condition, outputting emergency warning information corresponding to the emergency type.

[0024] The present invention first determines several types of emergencies, extracts the first element features of the collected raw data corresponding to each event type, and the extracted first element features contain features of the same emergency (historical event) from multiple angles and at different times. It is necessary to associate the features belonging to the same emergency with the corresponding emergency, that is, to achieve the above-mentioned fusion calculation, and then perform fusion calculation on the extracted first element features to obtain the first portrait feature. The first portrait feature is the emergency feature of the event presented by the real historical event of the corresponding type of emergency, and the emergency portrait model is constructed based on the first portrait feature.

[0025] After the emergency portrait model is constructed, real-time multi-source data of any emergency type is obtained through multi-source channels, and the second element features are extracted from them. The same method as above is used for fusion calculation to obtain several second portrait features. The aforementioned emergency portrait model performs in-depth analysis and processing on all the extracted second portrait features to obtain the emergency evolution trend corresponding to the emergency type. The emergency evolution trend describes whether the current actual features are developing in the direction of the actual occurrence of the emergency. When the emergency evolution trend meets the warning conditions, it means that based on the current real-time feature prediction, it is predicted that this type of emergency is likely to occur soon. At this time, a warning message is sent to the relevant ports for targeted response processing.

[0026] Among them, the types of emergencies include traffic incidents, crowd gathering incidents, natural disaster incidents, etc., which are not specifically limited in the present invention. The above-mentioned original data and multi-source data corresponding to different types of emergencies are also different. For example, the data corresponding to traffic incidents are mainly the monitoring data of the traffic monitoring system (such as traffic operation data collected by cameras, radars, etc.), the operation status data uploaded by floating vehicles (positioning, vehicle speed, etc.), etc., the data corresponding to crowd gathering incidents are mainly the terminal access device status (quantity, trend, etc.) of mobile communication devices, real-time video data taken by the city monitoring system, etc., and the data corresponding to natural disaster incidents include real-time detection data uploaded by monitoring sensors deployed at monitoring points, detection equipment such as drones, etc.

[0027] It should be noted that the emergency incident portrait model in the present invention is preferably built using a neural network. For the construction and training methods of the model, reference may be made to the existing patent CN114119057A, and the details will not be repeated here.

[0028] In some embodiments, the determining of several emergency event types and the corresponding first fine-grainedness includes: receiving several emergency event types set manually, evaluating the emergency degree of each emergency event type according to preset evaluation rules, and obtaining the emergency degree corresponding to each emergency event type; and determining the first fine-grainedness corresponding to each emergency event type according to the emergency degree.

[0029] In an embodiment of the present invention, relevant personnel can manually set a variety of emergency types according to monitoring requirements, such as the aforementioned traffic incidents, gathering of people incidents, natural disaster incidents, etc. After receiving the above-mentioned emergency types set manually, the preset evaluation rules are used to evaluate the degree of emergency of each emergency type. The degree of emergency is negatively correlated with the duration of the corresponding type of emergency from the appearance of signs of the incident to the occurrence of the incident. For example, the degree of emergency of traffic incidents is very high, while the degree of emergency of gathering of people incidents is much lower, and the degree of emergency of natural disaster incidents is also lower than that of traffic incidents. Finally, the first fine granularity corresponding to each emergency type is determined according to the degree of emergency. Among them, the first fine granularity refers to the time step when extracting the first element feature from the original data, that is, the first element feature is collected once in units of how many time intervals. Obviously, the smaller the time interval, the larger the corresponding first fine granularity, and vice versa.

[0030] Corresponding to the above, the preset evaluation rule can be the average value of the time from the appearance of signs of the emergency to the occurrence of the emergency in the historical events corresponding to the emergency type. The larger the average value, the lower the corresponding emergency degree, and vice versa. The above average value is obtained by statistically analyzing the historical data corresponding to each emergency type. In addition, the above evaluation rules can also be embedded in a specific evaluation model, such as based on a convolutional network.

[0031] In some embodiments, determining the first fine granularity corresponding to each of the emergency event types according to the emergency degree is specifically: ; In the above formula, For the first fine granularity, is a constant value, used to represent the regular value of the first fine-grainedness; The emergency degree corresponding to any of the emergency event types; is the adjustment parameter and is the preset value.

[0032] In an embodiment of the present invention, the first fine granularity and the degree of emergency corresponding to a specific emergency type are in a functional relationship in the above calculation formula. Based on the above calculation formula, it can be seen that when the degree of emergency of a certain emergency type is higher, that is, the average value of the time from the appearance of the event sign to the occurrence of the event is shorter, the first fine granularity is set to be smaller, that is, the first element feature of the original data is collected at a smaller time interval, so that the emergency portrait model constructed can perform more timely and accurate prediction and analysis of this type of emergency. On the contrary, when the degree of emergency of a certain emergency type is lower, that is, the average value of the time from the appearance of the event sign to the occurrence of the event is longer, the first fine granularity is set to be larger, that is, the first element feature of the original data is collected at a longer time interval, so that it is ensured that the emergency portrait model constructed can perform more timely and accurate prediction and analysis of this type of emergency, and at the same time, it can also reduce the model complexity and calculation load of the model for analyzing this type of emergency.

[0033] In some embodiments, the second element feature extraction of the real-time multi-source data includes: obtaining historical data of emergencies corresponding to the type of emergencies in the monitoring area, and deriving an event probability distribution function corresponding to the type of emergencies in the monitoring area based on statistics of the historical data of emergencies, wherein the event probability distribution function is used to characterize the time distribution of the probability of occurrence of events of a specific type of emergencies in the monitoring area; determining the probability of occurrence of an event corresponding to the current time period based on the event probability distribution function, and determining a second fine granularity based on the probability of occurrence of the event; and extracting the second element feature of the real-time multi-source data according to the second fine granularity.

[0034] In an embodiment of the present invention, emergency event historical data corresponding to different emergency event types in each monitoring area are retrieved, and these emergency event historical data record the occurrence data of a specific type of emergency event in the monitoring area, mainly the occurrence date and time data. By performing statistical analysis on these emergency event historical data, event probability distribution functions corresponding to different emergency event types can be obtained, and the event probability distribution function can be a normal distribution function, which is used to characterize the time distribution of the event occurrence probability of a specific emergency event type in the monitoring area, that is, the occurrence probability of this type of emergency event on different dates and in different time periods.

[0035] The probability of occurrence of the emergency event type corresponding to the current time period (including date) can be determined from the event probability distribution function, and the second fine granularity is determined according to the size of the event probability. The second element feature extraction is then performed on the real-time multi-source data collected above according to the second fine granularity.

[0036] Therefore, when performing feature extraction on real-time multi-source data, the present invention also adopts corresponding collection fine-grainedness in a targeted manner, thereby ensuring that the interval length of the collected second factor features is reasonable, which is conducive to obtaining more accurate evolution trends of emergencies in the future, and can effectively reduce the computational load when extracting the second factor features.

[0037] In some embodiments, the second fine granularity is determined according to the probability of occurrence of the event: ; In the above formula, For the second fine granularity, is a constant value, used to represent the regular value of the second finer granularity; The probability of occurrence of the emergency event type in the monitoring area in the current time period; is the adjustment parameter and is the preset value.

[0038] In the embodiment of the present invention, the second fine granularity and the probability of occurrence of the event in the current time period corresponding to the specific emergency type in the monitoring area are in a functional relationship in the above calculation formula. Based on the above calculation formula, it can be seen that when the probability of occurrence of the event of the specific type of emergency in the current time period is greater, the probability that the second element feature shows that the characteristics of the emergency are significant is greater, and the corresponding second fine granularity is set to be larger, that is, the second element feature of the real-time multi-source data is collected at a larger time interval, which can reduce the calculation load. When the probability of occurrence of the event of the specific type of emergency in the current time period is smaller, the probability that the second element feature shows that the characteristics of the emergency are significant is smaller, and the corresponding second fine granularity is set to be smaller, that is, the second element feature of the real-time multi-source data is collected at a smaller time interval, which can improve the prediction accuracy of the emergency in the low probability period. Such a setting can ensure that the interval length of the second element feature collected is as far as possible in a reasonable range, which is conducive to obtaining a more accurate evolution trend of the emergency in the future, and can effectively reduce the calculation load when the second element feature is extracted.

[0039] In some embodiments, Figure 2 As shown, all of the second portrait features are input into the emergency event portrait model to generate an emergency event evolution trend corresponding to each of the emergency event types, including: grouping all of the second portrait features in the order of collection and inputting them into the emergency event portrait model, the emergency event portrait model performs predictive analysis on each group of the second portrait features to obtain the corresponding event prediction probability of occurrence; curve fitting is performed on all of the event prediction probabilities in accordance with the above-mentioned order of collection to obtain an event prediction probability curve, which is the emergency event evolution trend; wherein the number of groups is determined according to the second fine-grainedness, and the number of groups is positively correlated with the second fine-grainedness.

[0040] In the embodiment of the present invention, the emergency event portrait model can perform predictive analysis on each set of input second portrait features, and output the corresponding predicted event probability. The time difference between each predicted event probability and the corresponding input second portrait feature is within the set time length, that is, the emergency event portrait model can predict the predicted event probability in the short future based on the real-time second portrait feature. Then, the predicted occurrence probability of all events is curve-fitted according to the above-mentioned collection sequence to obtain the predicted occurrence probability curve of the emergency event, which is the evolution trend of the emergency event.

[0041] When the probability of an emergency event, especially the latter one, is higher than the probability threshold, it can be determined that the emergency event of this type is about to occur, and warning information can be output at this time so that relevant personnel or equipment can handle it in time. For example, when it is determined that a certain area is about to have traffic congestion or crowds, a warning message can be issued, and traffic control personnel can go to the scene to clear the traffic in advance, or the traffic signal control system can adjust the regional signal in the area in advance, and management personnel can go to the scene of crowds to conduct guidance and management.

[0042] like Figure 3 As shown, an embodiment of the present invention also discloses an emergency warning system based on emergency portrait, the system comprising a first extraction module, a modeling module, a second extraction module, a prediction module, and an early warning module; the first extraction module is used to determine a number of emergency types and corresponding first fine-grainedness, collect raw data corresponding to each of the emergency types, and use the first fine-grainedness to extract first element features from the raw data; the modeling module is used to perform fusion calculation on the extracted first element features to obtain first portrait features, and construct an emergency portrait model based on the first target portrait features; the second extraction module is used to obtain real-time multi-source data corresponding to any of the emergency types through multi-source channels, perform second element feature extraction on the real-time multi-source data, perform fusion calculation on the extracted second element features to obtain a number of second portrait features; the prediction module is used to input all the second portrait features into the emergency portrait model to generate an emergency evolution trend corresponding to each of the emergency types; the early warning module is used to output emergency warning information corresponding to the emergency type when the emergency evolution trend meets the early warning condition.

[0043] An embodiment of the present invention further discloses an electronic device, comprising: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, characterized in that: the processor executes the computer program to implement the method described in the above embodiment.

[0044] An embodiment of the present invention further discloses a computer storage medium, wherein the computer storage medium stores a computer program, and is characterized in that: the computer program is executed by a processor to implement the method described in the above embodiment.

[0045] The embodiment of the present invention further discloses a computer program product. When the computer program product is run on a terminal, the terminal implements the method described in the above embodiment.

[0046] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0047] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0048] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0049] The above description is only a preferred embodiment of the present invention and is not intended to limit the protection scope of the present invention.

Claims

1. A method for early warning of emergencies based on emergency portraits, characterized in that , the method comprises the following steps: Determine a number of emergency event types and corresponding first fine granularities, collect raw data corresponding to each of the emergency event types, and use the first fine granularity to extract first element features from the raw data; wherein the first fine granularity refers to the time step when extracting the first element features from the raw data; Performing fusion calculation on the extracted first element features to obtain a first portrait feature, and constructing an emergency portrait model based on the first portrait feature; Acquire real-time multi-source data corresponding to any of the emergency event types through multi-source channels, extract second factor features from the real-time multi-source data, and perform fusion calculation on the extracted second factor features to obtain a plurality of second portrait features; Inputting all of the second portrait features into the emergency portrait model to generate emergency evolution trends corresponding to each of the emergency types; When the evolution trend of the emergency event meets the warning condition, outputting emergency event warning information corresponding to the emergency event type; Determining a plurality of emergency event types and corresponding first fine granularities includes: receiving a plurality of emergency event types manually set, evaluating the emergency degree of each emergency event type according to a preset evaluation rule, and obtaining the emergency degree corresponding to each emergency event type; and determining the first fine granularity corresponding to each emergency event type according to the emergency degree; Determining the first fine granularity corresponding to each of the emergency event types according to the emergency degree is specifically: ; In the above formula, For the first fine granularity, is a constant value, used to represent the regular value of the first fine-grainedness; The emergency degree corresponding to any of the emergency event types; Adjust the parameters to the preset values; Extracting the second factor feature from the real-time multi-source data includes: Acquire historical data of emergencies corresponding to the type of emergencies in the monitoring area, and obtain an event probability distribution function corresponding to the type of emergencies in the monitoring area based on the historical data of emergencies, wherein the event probability distribution function is used to characterize the time distribution of the occurrence probability of an event of a specific type of emergencies in the monitoring area; Determine the probability of an event corresponding to the current time period according to the event probability distribution function, and determine the second fine granularity according to the event probability; Extracting second element features from the real-time multi-source data according to the second fine granularity; Determine the second fine granularity according to the probability of occurrence of the event: ; In the above formula, For the second fine granularity, is a constant value, used to represent the regular value of the second finer granularity; The probability of occurrence of the emergency event type in the monitoring area in the current time period; is the adjustment parameter and is the preset value.

2. The method for early warning of an emergency event based on emergency event portrait according to claim 1, characterized in that: Inputting all the second portrait features into the emergency portrait model to generate emergency evolution trends corresponding to each of the emergency types, including: Grouping all the second portrait features in the order of collection and inputting them into the emergency event portrait model, the emergency event portrait model performs prediction analysis on each group of the second portrait features to obtain the corresponding event prediction probability; The predicted occurrence probabilities of all the events are curve-fitted according to the above-mentioned collection order to obtain the event prediction probability curve, which is the evolution trend of the sudden events; wherein the number of groups is determined according to the second fine granularity, and the number of groups is positively correlated with the second fine granularity.

3. An emergency warning system based on emergency portrait, the system is based on the method of claim 1 or 2, the system comprises a first extraction module, a modeling module, a second extraction module, a prediction module, and an early warning module; characterized in that: The first extraction module is used to determine a number of emergency event types and corresponding first fine granularities, collect raw data corresponding to each of the emergency event types, and use the first fine granularity to extract first element features from the raw data; wherein the first fine granularity refers to a time step when extracting the first element features from the raw data; The modeling module is used to perform fusion calculation on the extracted first element features to obtain a first portrait feature, and to construct an emergency portrait model based on the first portrait feature; The second extraction module is used to obtain real-time multi-source data corresponding to any of the emergency types through multi-source channels, extract second element features from the real-time multi-source data, and perform fusion calculation on the extracted second element features to obtain a plurality of second portrait features; The prediction module is used to input all the second portrait features into the emergency portrait model to generate an emergency evolution trend corresponding to each of the emergency types; The warning module is used to output emergency warning information corresponding to the emergency type when the emergency evolution trend meets the warning condition.

4. An electronic device comprising: At least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the processor executes the computer program to implement the method according to claim 1 or 2.

5. A computer storage medium storing a computer program, characterized in that: The computer program is executed by a processor to implement the method according to claim 1 or 2.

6. A computer program product comprising a computer program stored on a non-transitory computer readable medium, characterized in that: When the computer program is executed by a processor, the method according to claim 1 or 2 is implemented.

Images