Method, device, computer program product and electronic equipment for determining regional risk

Abstract

The application discloses a method and device for determining regional risk, a computer program product and an electronic device. It relates to the field of public safety or other related fields. The method comprises the following steps: obtaining risk prediction requirements of a target region, determining target data types according to the risk prediction requirements, wherein the target data types are used to represent the data types required when predicting the risk of the target region; obtaining initial data conforming to the target data types from different data sources to obtain M initial data, and performing data fusion on the M initial data to obtain N target fusion data, wherein M is greater than N, and M and N are positive integers; inputting the N target fusion data into a risk prediction model to output a risk prediction result, wherein after obtaining the risk prediction result, an early warning strategy is determined based on the risk prediction result. Through the application, the problem of insufficient accuracy and timeliness of identifying and warning the risk of a region in related technologies is solved.

Description

Technical Field

[0001] This application relates to the field of public safety or other related fields, and more specifically, to a method, apparatus, computer program product, and electronic device for determining regional risks. Background Technology

[0002] With the acceleration of global urbanization, the continuous expansion of city size, and the rapid concentration of population, the complexity of urban operating systems has increased significantly. To ensure the normal operation of cities and the safety of residents, urban risk perception and early warning have become crucial. Urban risks include various types, such as natural disasters and public safety incidents. These risks not only affect the stability of urban infrastructure but also directly relate to the safety of residents' lives and property. Currently, urban risk perception systems mainly rely on the analysis of single data sources, such as meteorological data, video surveillance, and social media data. While these technologies can effectively identify and warn of some risks in specific scenarios, their limitations cannot be ignored.

[0003] First, analysis of a single data source cannot capture the full picture of urban risks and lacks a comprehensive reflection of the complexity and diversity of risks, which can easily lead to incomplete risk identification and increase the risk of false alarms and missed alarms. Since information from different data sources is often isolated, it can easily lead to fragmented risk perception and make it difficult to form a comprehensive urban risk cognition framework, thereby affecting the accuracy and timeliness of risk warnings.

[0004] Furthermore, with the increase in urban scale and the diversification of risk types, the demand for risk perception systems is increasing. However, the relevant technologies are clearly insufficient in terms of data processing and model selection, making it difficult to meet residents' urgent needs for risk perception, especially when dealing with large-scale, sudden risk events.

[0005] There is currently no effective solution to the problem of insufficient accuracy and timeliness in identifying and warning of regional risks in related technologies. Summary of the Invention

[0006] The main objective of this application is to provide a method, apparatus, computer program product, and electronic device for determining regional risks, in order to solve the problems of insufficient accuracy and timeliness in the identification and early warning of regional risks in related technologies.

[0007] To achieve the above objectives, according to one aspect of this application, a method for determining regional risk is provided. The method includes: obtaining risk prediction requirements for a target region; determining a target data type based on the risk prediction requirements, wherein the target data type characterizes the data type required for predicting the risk of the target region; obtaining initial data conforming to the target data type from different data sources to obtain M initial data points; fusing the M initial data points to obtain N target fused data points, wherein M is greater than N, and M and N are positive integers; inputting the N target fused data points into a risk prediction model and outputting a risk prediction result; wherein, after obtaining the risk prediction result, a warning strategy is determined based on the risk prediction result.

[0008] Furthermore, initial data conforming to the target data type is obtained from different data sources to obtain M initial data points. This includes: determining multiple data sources based on the target data type, where each data source refers to a device in the target area that collects M initial data points; determining screening conditions based on risk prediction requirements; extracting K data points to be processed from the data sources based on the screening conditions; and preprocessing the K data points to be processed to obtain M initial data points, where K is greater than or equal to M and K is a positive integer.

[0009] Furthermore, the data fusion of M initial data to obtain N target fused data includes: classifying the M initial data according to the scene features of each initial data to obtain N sets of initial modal data, where each set of initial modal data is associated with a scene type; performing vector transformation on the initial features associated with the initial data in each of the N sets of initial modal data to obtain N sets of feature vectors; determining the learning weights of the initial features of each of the M initial data to obtain M learning weights; and for a set of initial modal data, performing a weighted calculation on a set of feature vectors associated with the set of initial modal data and the learning weights of the initial features associated with the set of initial modal data to obtain one set of target fused data.

[0010] Furthermore, the M initial data are fused to obtain N target fused data, including: classifying the M initial data according to the scene features of each initial data to obtain N sets of initial modal data, where each set of initial modal data is associated with a scene type; obtaining Y scene categories, determining the predicted probability of each set of initial modal data belonging to the Y scene categories, to obtain N sets of predicted probability sets, where each set of predicted probability sets includes Y predicted probabilities, where Y is a positive integer; determining the prediction weight of each set of predicted probability sets to obtain N prediction weights, and performing a weighted calculation on each prediction weight and each set of predicted probability sets to obtain N target fused data.

[0011] Furthermore, the process of fusing M initial data to obtain N target fused data includes: performing vector transformation on the initial features of the M initial data to obtain M feature vectors; for each feature vector, using a probability allocation function to calculate the normalization factor of the feature vector with each of the M-1 feature vectors, obtaining M-1 normalization factors, and using the M-1 normalization factors to determine the overlap data of the feature vector; for each feature vector, determining the overlap data of the feature vector with each of the M-1 feature vectors, obtaining M-1 overlap data; constructing a feature matrix based on the M overlap data and the M overlap data, and determining the N target fused data based on the feature matrix.

[0012] Furthermore, the early warning strategy is determined as follows: risk features are extracted from the risk prediction results, where the risk features are used to characterize the degree of correlation between each target fusion data and P risk categories, where P is a positive integer; the levels of the P risk categories are determined according to the risk features to obtain P risk levels, and an early warning strategy is generated from the P risk levels.

[0013] Furthermore, the risk prediction model is trained as follows: Q historical data points over a historical time period are obtained, and an initial risk prediction model is obtained, where Q is a positive integer; the initial risk prediction model is trained unsupervised using the Q historical data points to obtain a first prediction model; P risk categories are obtained, and the Q historical data points are labeled with the P risk categories to obtain Q historical category data points, and the first prediction model is trained in a supervised manner using the Q historical feature data points to obtain a second prediction model; knowledge distillation is performed on the second prediction model to obtain the risk prediction model.

[0014] To achieve the above objectives, according to another aspect of this application, a device for determining regional risk is provided. The device includes: a first acquisition unit, configured to acquire risk prediction requirements for a target region and determine a target data type based on the risk prediction requirements, wherein the target data type characterizes the data type required for predicting the risk of the target region; a second acquisition unit, configured to acquire initial data conforming to the target data type from different data sources, obtaining M initial data points, and fusing the M initial data points to obtain N target fused data points, wherein M is greater than N, and M and N are positive integers; and an input unit, configured to input the N target fused data points into a risk prediction model and output a risk prediction result, wherein after obtaining the risk prediction result, an early warning strategy is determined based on the risk prediction result.

[0015] According to another aspect of the present invention, a computer storage medium is also provided for storing a program, wherein the program, when running, controls the device where the computer storage medium is located to execute a method for determining regional risk.

[0016] According to another aspect of the present invention, an electronic device is also provided, comprising one or more processors and a memory; the memory stores computer-readable instructions, and the processor is configured to execute the computer-readable instructions, wherein the computer-readable instructions, when executed, perform a method for determining regional risk.

[0017] According to another aspect of the present invention, a computer program product is also provided, including a computer program that, when executed by a processor, performs a method for determining regional risk.

[0018] This application employs the following steps: First, it obtains the risk prediction needs of the target region; second, it determines the target data type based on these needs, where the target data type characterizes the data type required for predicting risks in the target region; third, it obtains initial data conforming to the target data type from different data sources, resulting in M ​​initial data sets, and then fuses these M initial data sets to obtain N target fused data sets, where M is greater than N and M and N are positive integers; fourth, it inputs the N target fused data sets into a risk prediction model and outputs risk prediction results. After obtaining the risk prediction results, it determines an early warning strategy based on these results. This addresses the issues of insufficient accuracy and timeliness in identifying and issuing early warnings of regional risks in related technologies. By identifying the risk prediction needs of the target region, obtaining the target data type, obtaining initial data based on the data type, fusing the initial data to obtain the target fused data, and finally processing the target fused data using a risk prediction model to obtain the risk prediction results, it achieves the effect of improving the accuracy and efficiency of identifying and issuing early warnings of regional risks. Attached Figure Description

[0019] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0020] Figure 1 This is a flowchart of a method for determining regional risk according to an embodiment of this application;

[0021] Figure 2 This is a schematic diagram of a training method for a risk prediction model provided according to an embodiment of this application;

[0022] Figure 3 This is a schematic diagram of an urban risk perception platform provided according to an embodiment of this application;

[0023] Figure 4 This is a schematic diagram of an urban risk perception system provided according to an embodiment of this application;

[0024] Figure 5This is a schematic diagram of a regional risk determination device provided according to an embodiment of this application;

[0025] Figure 6 This is a schematic diagram of an electronic device provided according to an embodiment of this application. Detailed Implementation

[0026] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0027] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0028] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0029] It should be noted that all information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for display, data used for analysis, etc.) involved in this disclosure are information and data authorized by the user or fully authorized by all parties. For example, this system has an interface with relevant users or organizations. Before obtaining relevant information, it is necessary to send an acquisition request to the aforementioned user or organization through the interface, and obtain the relevant information after receiving consent information from the aforementioned user or organization.

[0030] It should be noted that the information collected in this application is information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of the relevant data all comply with the relevant laws, regulations and standards of the relevant regions, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation access points for users to choose to authorize use or refuse use.

[0031] The present invention will now be described in conjunction with preferred implementation steps. Figure 1 This is a flowchart of a method for determining regional risk according to an embodiment of this application, such as... Figure 1 As shown, the method includes the following steps:

[0032] Step S101: Obtain the risk prediction requirements of the target area, and determine the target data type based on the risk prediction requirements. The target data type is used to characterize the data type required when predicting the risk of the target area.

[0033] Specifically, risk prediction needs can include regional high temperature prediction needs, traffic condition prediction needs, and rainstorm prediction needs. The target area can refer to the region (such as a city) that needs risk prediction. In order to improve the accuracy and timeliness of risk prediction, after obtaining the risk prediction needs of the target area, the risk prediction needs are first identified, and then the environmental and urban infrastructure information of the target area is determined based on the needs. For example, when the target area is a coastal city, the risk prediction needs of the area can be the prediction needs of risks such as rainstorms, sea-level rise, and drainage system effectiveness; when the target area is an earthquake-prone area, the risk prediction needs can be the prediction needs of geological changes.

[0034] Furthermore, after extracting relevant information from the aforementioned risk prediction needs, the target data type required for predicting regional risks can be determined based on the extracted information. For example, after identifying the risk prediction needs of coastal cities, the determined target data type can be meteorological data (such as real-time temperature and humidity, wind speed and direction, air pressure, precipitation, and other meteorological parameters, as well as long-term meteorological trends and seasonal variation data); after identifying the risk prediction needs of earthquake-prone cities, the determined target data type can be geospatial data (such as topography, river distribution, building density, population distribution, and other information).

[0035] Step S102: Obtain initial data that conforms to the target data type from different data sources to obtain M initial data, and fuse the M initial data to obtain N target fused data, where M is greater than N and M and N are positive integers.

[0036] Specifically, the data source can be sensor devices such as earthquake monitoring sensors, water level sensors, smoke sensors, and air quality monitors, or it can be a weather station or a satellite remote sensing system. After determining the target data type, initial data can be extracted from multiple data sources based on these data types. For example, when the target data type is meteorological data, real-time and historical meteorological data, such as temperature, humidity, precipitation, wind speed, and wind direction, can be obtained from weather stations or satellite remote sensing systems. When the target data type is video data, multiple video data can be obtained through camera devices in the city, and then video analysis algorithms can be used to identify objects, behaviors, and environmental changes, thereby identifying public safety events such as traffic congestion.

[0037] In addition, the data source can also be a database storing historical time periods, which may include historical disaster records, emergency response cases, urban operation data, etc. Through this data, risk prediction models can be trained, enabling them to learn and master risk patterns and development trends.

[0038] It should be noted that after obtaining the initial data stored in the data source, multimodal fusion processing can be performed on these data to obtain a more comprehensive, accurate and valuable information set. Among these methods, feature-level fusion, decision-level fusion or hybrid fusion can be used to process the data, which can not only improve the comprehensive utilization rate of the data, but also provide a more comprehensive and in-depth information foundation for subsequent risk identification and assessment.

[0039] Step S103: Input the fused data of N targets into the risk prediction model and output the risk prediction result. After obtaining the risk prediction result, determine the early warning strategy based on the risk prediction result.

[0040] Specifically, risk prediction models can be deep learning models that can process, fuse, and understand different types of data (such as text, images, audio, video, etc.). By learning the common semantics of information from various forms, they can achieve interaction between different modalities. Risk prediction models can be convolutional neural network models, long short-term memory network models, etc.

[0041] Since the processed target fusion data contains multi-faceted descriptions of urban risks, such as meteorological parameters and image features, this data can be used as input to a risk prediction model, ensuring that the model can comprehensively understand the complexity and diversity of potential risks. After the risk prediction model obtains the aforementioned target fusion data, it can analyze the data to predict possible future risk events and their impact, thereby outputting risk prediction results. These results, representing the probability distribution of one or more risk events, can include information such as event type, probability of occurrence, and expected scope of impact.

[0042] Furthermore, the risk prediction results can be analyzed to assess the urgency and severity of the risk events, and then different early warning strategies can be determined. For example, when the risk prediction results indicate that there is a high probability of flooding in a certain area within the next 24 hours, different early warning strategies can be formulated based on the possible inundation depth and scope of impact, and the early warning information can be quickly disseminated to residents in the affected areas through various channels such as SMS platforms, social media, radio, and television.

[0043] The method for determining regional risk provided in this application involves acquiring the risk prediction needs of a target region, determining the target data type based on these needs (where the target data type characterizes the data type required for predicting the risk of the target region), acquiring initial data conforming to the target data type from different data sources to obtain M initial data sets, and fusing these M initial data sets to obtain N target fused data sets, where M is greater than N and M and N are positive integers. The N target fused data sets are then input into a risk prediction model to output risk prediction results. After obtaining the risk prediction results, an early warning strategy is determined based on these results. This method addresses the issues of insufficient accuracy and timeliness in identifying and issuing early warnings of regional risks in related technologies. By identifying the risk prediction needs of the target region, obtaining the target data type, acquiring initial data based on the data type, fusing the initial data to obtain the target fused data, and finally processing the target fused data using a risk prediction model to obtain the risk prediction results, the method improves the accuracy and efficiency of identifying and issuing early warnings of regional risks.

[0044] To improve the accuracy and efficiency of risk identification and early warning, optionally, in the method for determining regional risks provided in this application embodiment, obtaining initial data conforming to the target data type from different data sources to obtain M initial data includes: determining multiple data sources according to the target data type, wherein each data source refers to a device in the target region that collects M initial data; determining screening conditions according to risk prediction needs; extracting K data to be processed from the data sources according to the screening conditions; and preprocessing the K data to be processed to obtain M initial data, wherein K is greater than or equal to M and K is a positive integer.

[0045] Specifically, when collecting initial data using a target data type, the first step is to determine the corresponding data source based on the defined target data type. For example, when the target data type is geographic environmental data, the data source can be environmental data sensors distributed in different locations within the region, such as IoT devices like earthquake monitoring sensors, water level sensors, smoke sensors, and air quality monitors used for early warning of structural safety and environmental changes. When the target data type is meteorological data, the data source can be a meteorological data source, such as a meteorological station or satellite remote sensing system capable of collecting meteorological parameters like temperature, humidity, precipitation, wind speed, and atmospheric pressure. When the target data type is video data, the data source can be a network of surveillance cameras in the city used to analyze changes in objects, behaviors, and the environment. The data source can also be a historical data source associated with a historical database, which can store meteorological records, disaster cases, and emergency response records for historical time periods, enabling model training and risk pattern recognition.

[0046] Furthermore, to improve the targeting and efficiency of data processing, after determining the data source, filtering conditions can be established based on specific risk prediction needs, such as risk type, time range, and geographical region. Then, initial data relevant to risk prediction can be precisely extracted using these filtering conditions. For example, when predicting typhoon risk, filtering conditions could be meteorological data within a specific time range; when monitoring public safety incidents, filtering conditions could be content related to specific geographical locations and times. After extracting multiple data points from the data source using filtering conditions, preprocessing operations such as data cleaning can be performed on these data to obtain initial data. This embodiment, by extracting initial data from the data source using filtering conditions, can collect information directly related to risk prediction, covering multiple dimensions of risk events. This significantly improves the depth and breadth of risk prediction, reduces false positives and false negatives that may arise from a single data source, and improves the accuracy of model predictions.

[0047] Optionally, in the method for determining regional risk provided in this application embodiment, fusing M initial data to obtain N target fused data includes: classifying the M initial data according to the scene features of each initial data to obtain N sets of initial modal data, wherein each set of initial modal data is associated with a scene type; performing vector transformation on the initial features associated with the initial data in each set of initial modal data in the N sets of initial modal data to obtain N sets of feature vectors; determining the learning weights of the initial features of each initial data in the M initial data to obtain M learning weights; and for a set of initial modal data, performing a weighted calculation on a set of feature vectors associated with the set of initial modal data and the learning weights of the initial features associated with the set of initial modal data to obtain a target fused data.

[0048] Specifically, after determining the initial data using screening criteria, to help the risk prediction model output more comprehensive risk prediction results, the initial data can be fused. First, the scene characteristics of each initial data point can be identified. For example, meteorological data corresponds to weather scenes, video surveillance data corresponds to public safety incident scenes, and social media data corresponds to social change scenes, etc. Then, based on these scene characteristics, the initial data can be classified into different scenarios, resulting in multiple sets of initial modal data corresponding to different scenario types, such as initial modal data for natural disaster scenarios and initial modal data for public safety scenarios.

[0049] Furthermore, after obtaining the aforementioned initial modal data, feature extraction can be performed on each initial data point in each set of initial modal data, and the extracted initial features can be converted into feature vectors. Simultaneously, the learning weight of each feature vector is determined. For example, in an earthquake early warning scenario, since the geographical feature vectors of earthquake monitoring sensors can directly reflect the physical characteristics of earthquake activity, these vectors can be associated with higher learning weights. Finally, all feature vectors and their corresponding learning weights are weighted and calculated, i.e., the target fusion data corresponding to each set of initial modal data is calculated using the following formula. Where, α i Representative eigenvector The learning weights are denoted by , where i represents the i-th feature vector. This embodiment determines the learning weights of the feature vectors of each initial data point, and then uses feature-level fusion to determine the target fused data. This allows for effective combination based on scene characteristics, effectively avoiding the information bias and limitations of analyzing a single data source. This helps the model more accurately capture risk-related features, improving the accuracy and timeliness of early warnings.

[0050] Data fusion methods include various approaches. Optionally, in the method for determining regional risk provided in this application embodiment, fusing M initial data to obtain N target fused data includes: classifying the M initial data according to the scene characteristics of each initial data to obtain N sets of initial modal data, wherein each set of initial modal data is associated with a scene type; obtaining Y scene categories, determining the predicted probability that each set of initial modal data belongs to the Y scene categories, obtaining N sets of predicted probability sets, wherein each set of predicted probability sets includes Y predicted probabilities, where Y is a positive integer; determining the prediction weight of each set of predicted probability sets, obtaining N prediction weights, and performing a weighted calculation on each prediction weight and each set of predicted probability sets to obtain N target fused data.

[0051] Specifically, to improve the prediction accuracy of risk prediction models, decision-level fusion can be used for data processing. First, after classifying the scene features based on each initial data point and obtaining multiple sets of initial modal data, the probability of each set of initial modal data belonging to each scene category can be calculated based on the scene type of the initial modal data. This yields a corresponding set of prediction probabilities, which involves calculating the similarity between scene type and each scene category to determine the prediction probability. Further, after determining the prediction weights based on scene categories, the target fused data can be calculated using the prediction weights and the prediction probability set. The target fused data P can then be calculated using the following formula. fused (y): Among them, P i (y) represents the predicted probability that the i-th initial modal data belongs to the y-th scene category, β i This represents the prediction weight of the prediction probability set corresponding to each set of initial modal data. This embodiment utilizes decision-level fusion to calculate the target fusion data, which can cover a wide range of risk scenario types. At the same time, by calculating the prediction probability, it provides quantitative risk assessment, significantly improving the comprehensiveness and accuracy of risk identification.

[0052] Besides using feature-level fusion and decision-level fusion for data processing, data fusion can also be achieved based on multiple fusion methods. Optionally, in the method for determining regional risk provided in this application embodiment, fusing M initial data to obtain N target fused data includes: performing vector transformation on the initial features of the M initial data to obtain M feature vectors; for each feature vector, using a probability allocation function to calculate the normalization factor of the feature vector with M-1 feature vectors respectively, obtaining M-1 normalization factors, and using the M-1 normalization factors to determine the compatibility data of the feature vector; for each feature vector, determining the overlap data of the feature vector with M-1 feature vectors respectively, obtaining M-1 overlap data; constructing a feature matrix based on the M overlap data and the M compatibility data, and determining N target fused data based on the feature matrix.

[0053] Specifically, when using a multi-fusion approach for data processing, features can first be extracted from all the acquired initial data to obtain corresponding initial features. Then, these initial features can be vectorized to obtain corresponding feature vectors. For example, the initial features can be visual features such as objects and behaviors extracted from image data, semantic features such as keywords and themes extracted from text data, or statistical features such as numerical changes and trends obtained from sensor data.

[0054] Furthermore, the probability allocation function (i.e., the mass function) in the Dempster-Shafer (DS) theory can be used to calculate the normalization factor for each eigenvector. That is, the normalization factor of the i-th eigenvector and the j-th eigenvector can be calculated using the mass function mentioned above. Here, the DS theory can handle uncertainty and incomplete information, and the mass function can represent the degree of confidence in a specific proposition.

[0055] Then, based on the normalization factor mentioned above, the degree of conflict or inconsistency between the information represented by each pair of feature vectors is calculated. That is, the reciprocity data of each feature vector is calculated by the following formula: Conf(i,j)=-log2(1-C), where C represents the normalization factor, and i and j represent the i-th and j-th feature vectors, respectively.

[0056] Then, using the reciprocity data of every two eigenvectors, the reciprocity data of a single eigenvector with all other eigenvectors is calculated: Where t represents the total number of data standard types, and Conf(i,k) represents the similarity data between the i-th and k-th feature vectors. Then, based on the probability allocation function, a coincidence function representing the degree of similarity or consistency is constructed, that is, the coincidence data of the i-th feature vector relative to all other feature vectors is determined according to the following formula: Where τ represents the boundary value for determining whether feature vectors overlap, and d ik This represents the distance between two eigenvectors determined by the mass function.

[0057] Furthermore, the overlap and reciprocity data of each feature vector are normalized to obtain the importance function: Where ε0 represents the normalization threshold. The feature matrix {a} is constructed based on the above formula. mn}, where the element a in the matrix mn This indicates the importance of data standard m relative to data standard n:

[0058] Finally, the target fused data S is determined using the aforementioned matrix. m : This embodiment calculates target fusion data through multiple fusion methods, which can quantitatively assess the correlation and complementarity between different data sources, effectively improve the fusion effect of multimodal data, enhance the understanding of complex risk scenarios, avoid false alarms and false negatives that may be caused by single data source analysis, and significantly improve the accuracy and reliability of risk identification.

[0059] Optionally, in the method for determining regional risks provided in this application embodiment, the early warning strategy is determined in the following manner: extracting risk features from the risk prediction results, wherein the risk features are used to characterize the degree of correlation between each target fusion data and P risk categories, where P is a positive integer; determining the level of the P risk categories based on the risk features to obtain P risk levels, and generating an early warning strategy from the P risk levels.

[0060] After the risk prediction model outputs the risk prediction results, the corresponding early warning strategy can be determined based on the risk prediction results. Specifically, firstly, multiple preset risk categories, such as natural disasters, can be obtained, and then risk characteristics related to the above risk categories can be extracted from the risk prediction results. For example, risk characteristics can include the frequency, intensity, scope of impact, and time pattern of risk events that can characterize the degree of correlation with different risk categories.

[0061] Furthermore, risk levels can be determined based on risk characteristics, and corresponding early warning strategies can be generated based on risk categories and risk levels. These early warning strategies can include the content of the early warning information, the method of early warning, the target group, and the timing of the early warning, ensuring that the early warning information is received by the right people at the right time. This embodiment, by utilizing risk prediction results to determine early warning strategies, helps optimize resource allocation and emergency response strategies, enhances the city's decision support capabilities in the face of various risks, and ensures rapid response and effective handling when risk events occur.

[0062] Optionally, in the method for determining regional risk provided in this application embodiment, the risk prediction model is trained in the following manner: acquiring Q historical data points over a historical time period and acquiring an initial risk prediction model, where Q is a positive integer; performing unsupervised training on the initial risk prediction model using the Q historical data points to obtain a first prediction model; acquiring P risk categories, labeling the Q historical data points with the P risk categories to obtain Q historical category data, and performing supervised training on the first prediction model using the Q historical feature data to obtain a second prediction model; and performing knowledge distillation on the second prediction model to obtain a risk prediction model.

[0063] Specifically, in order to improve the accuracy of risk prediction, it is necessary to optimize and train the risk prediction model. Figure 2 This is a schematic diagram of the training method for the risk prediction model provided in the embodiments of this application, such as... Figure 2 As shown, the process begins by collecting and integrating multiple historical data sets from past time periods. These historical data sets can include historical weather records, public safety incident records, and so on. Then, these historical data sets undergo data fusion processing using strategies such as large-scale model processing, feature-level fusion, decision-level fusion, or hybrid fusion to obtain multiple historical fused data sets.

[0064] Furthermore, after obtaining the initial risk prediction model, unsupervised training can be performed using the aforementioned historical fusion data to obtain a usable pre-trained model (i.e., the first prediction model) with preliminary risk identification and prediction capabilities. This model reduces reliance on large amounts of data, laying a solid foundation for subsequent supervised training. After obtaining the risk categories, the aforementioned historical fusion data can be labeled based on the risk categories to obtain historical category data representing different risk scenarios. Then, supervised training of the first prediction model can be performed using this historical category data to obtain an application-adapted model (i.e., the second prediction model). It should be noted that during this supervised training phase, the model can learn specific features closely related to the risk categories, improving the accuracy of risk prediction.

[0065] Furthermore, to improve prediction accuracy, the second prediction model can be further optimized through project refinement to obtain a real-time risk prediction model. Finally, knowledge distillation is performed on the real-time risk prediction model, that is, by transferring the decision knowledge of the large model to the small model, a risk prediction model with better computational efficiency and generalization ability (i.e., a high-precision risk prediction small model) is obtained.

[0066] It should be noted that when optimizing the above model, model parameters can be adjusted through loss functions, learning rates, regularization, etc., to ensure that the model can comprehensively capture the complex relationships between data. Regarding the learning rate, it is first necessary to determine a learning rate that controls the distance the agent moves along the negative gradient direction at each step. The learning rate can be calculated using the basic gradient descent formula: Where, θ t It is the current parameter, θ t+1 These are the updated parameters, J(θ) t ) is the loss function value under the current parameters. It is the gradient of the loss function with respect to the parameters, and η represents the learning rate.

[0067] Since the model needs to minimize the objective function value L(γ) to complete the optimization, the minimization of the objective function value L(γ) can be expressed as: minL(γ), γ={γ1,γ2,…,γ n}, and based on Lagrange's theorem and the projection method, the update strategy is derived as: γ l+1 ←γ l +ρ t D t , where ρ t The step size, D, is determined by a linear search method. t It is the vector of the gradient descent direction.

[0068] Furthermore, regularization methods can reduce model overfitting by adding a penalty term to the loss function. The principle of L1 regularization is expressed as follows: Where J(θ) is the original loss function, λ is the regularization coefficient used to control the strength of the regularization term, and n is the number of parameters. L1 regularization helps to produce a sparse weight matrix, i.e., many weights become zero; the loss function can be expressed as: This embodiment optimizes the risk prediction model through unsupervised pre-training and supervised fine-tuning, which helps the model learn the broad characteristics and specific patterns of risk events, significantly improving the accuracy and stability of risk prediction, while significantly reducing the model's computational resource requirements and improving the efficiency of real-time data processing and risk warning.

[0069] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0070] This application also provides a system for determining regional risks. Figure 3 This is a schematic diagram of an urban risk perception platform provided according to an embodiment of this application. Figure 4 This is a schematic diagram of an urban risk perception system provided according to an embodiment of this application, such as... Figure 3 , Figure 4 As shown, the platform includes: an urban risk perception system, a business application module, and a foundation support module. The urban risk perception system includes a data acquisition layer, a data processing layer, a model training layer, and an application layer. The foundation support module includes a hardware equipment sub-module and an algorithm capability sub-module. The hardware equipment sub-module includes a CPU, a storage server, an algorithm GPU, and a cloud server. The algorithm capability sub-module includes a multimodal large model. The business application module includes a platform application layer, an algorithm application layer, and an auxiliary application layer.

[0071] It should be noted that the platform's application layer enables manual review, intelligent grading, algorithmic feedback, risk classification, alarm push notifications, and early warning dissemination. This means it can identify potential public safety incidents based on fused data, assess their urgency and impact, generate targeted early warning information, and disseminate warnings through multiple channels. The algorithm application layer enables predictions of rising water levels, heavy rain and flooding, air quality, earthquake disasters, fire warnings, crowd gatherings, intelligent transportation, armed attacks, and fights. The auxiliary application layer enables departmental management, job access control, process management, video sharing management, data control and processing, and parameter settings. This provides management functions for different emergency departments, including resource allocation and personnel scheduling, ensuring efficient and orderly emergency response. Process management defines standard processes for generating, reviewing, and disseminating early warning information, ensuring accuracy and reliability. Job access control sets different platform access permissions based on user roles and responsibilities, protecting data security and privacy. Parameter settings allow users to adjust model parameters and early warning strategies to adapt to the public safety needs of specific areas.

[0072] The hardware infrastructure submodule includes CPUs and GPUs that provide computing resources to support the training of deep learning models and real-time data processing; storage servers that store historical data and model training results, ensuring data integrity and accessibility; and cloud servers that provide elastic computing and storage services, ensuring the platform's high availability and scalability. The algorithm capabilities submodule includes multimodal large models that enable efficient integration and analysis of cross-modal data.

[0073] When predicting regional risks, it is first necessary to determine the target data type required for the prediction. At this time, the data acquisition layer can collect multiple initial data that conform to the target data type, and the data processing layer can perform data cleaning, labeling, feature extraction and data fusion processing on the collected data to obtain the target fused data.

[0074] Furthermore, after training the risk prediction model using the model training layer in the aforementioned system, the target fusion data is input into the trained risk prediction model. The risk prediction model outputs risk prediction results, which are then evaluated by the application layer. Specifically, the algorithm application layer processes the risk prediction results through risk identification, risk assessment, alarm generation, and algorithm feedback to obtain an alarm strategy. Finally, the platform application layer publishes the alarm strategy to city residents through alarm processing, alarm push notifications, and early warning dissemination.

[0075] This embodiment utilizes an urban risk perception platform to predict regional risks, enabling a more comprehensive and in-depth understanding of urban operations and potential risks. When a risk event occurs, it can quickly provide accurate risk information, accelerate emergency response, reduce losses, and improve the effectiveness of emergency handling.

[0076] This application also provides a regional risk determination apparatus. It should be noted that the regional risk determination apparatus of this application can be used to execute the regional risk determination method provided in this application. The regional risk determination apparatus provided in this application will be described below.

[0077] Figure 5 This is a schematic diagram of a regional risk determination device provided according to an embodiment of this application, such as... Figure 5 As shown, the device includes: a first acquisition unit 50, a second acquisition unit 51, and an input unit 52.

[0078] The first acquisition unit 50 is used to acquire the risk prediction needs of the target area and determine the target data type based on the risk prediction needs. The target data type is used to characterize the data type required when predicting the risk of the target area.

[0079] The second acquisition unit 51 is used to acquire initial data that conforms to the target data type from different data sources, obtain M initial data, and fuse the M initial data to obtain N target fused data, where M is greater than N and M and N are positive integers;

[0080] Input unit 52 is used to input the fused data of N targets into the risk prediction model and output the risk prediction result. After obtaining the risk prediction result, an early warning strategy is determined based on the risk prediction result.

[0081] The regional risk determination apparatus provided in this application embodiment acquires the risk prediction needs of a target region through a first acquisition unit 50, and determines the target data type based on the risk prediction needs. The target data type is used to characterize the data type required when predicting the risk of the target region. A second acquisition unit 51 acquires initial data that conforms to the target data type from different data sources to obtain M initial data, and fuses the M initial data to obtain N target fused data, where M is greater than N and M and N are positive integers. An input unit 52 inputs the N target fused data into a risk prediction model and outputs the risk prediction result. After obtaining the risk prediction result, an early warning strategy is determined based on the risk prediction result. This solves the problem of insufficient accuracy and timeliness in the identification and early warning of regional risks in related technologies. By identifying the risk prediction needs of the target region, obtaining the target data type, acquiring initial data based on the data type, fusing the initial data to obtain the target fused data, and finally processing the target fused data using the risk prediction model to obtain the risk prediction result, the apparatus achieves the effect of improving the accuracy and efficiency of regional risk identification and early warning.

[0082] Optionally, in the regional risk determination device provided in the embodiments of this application, the second acquisition unit 51 includes: a first determination module, used to determine multiple data sources according to the target data type, wherein each data source refers to a device that collects M initial data in the target region; and a second determination module, used to determine screening conditions according to risk prediction requirements, extract K data to be processed from the data sources according to the screening conditions, and preprocess the K data to be processed to obtain M initial data, wherein K is greater than or equal to M and K is a positive integer.

[0083] Optionally, in the regional risk determination device provided in this application embodiment, the second acquisition unit 51 includes: a first classification module, used to perform modal classification on M initial data according to the scene features of each initial data to obtain N sets of initial modal data, wherein each set of initial modal data is associated with a scene type; a first conversion module, used to perform vector conversion on the initial features associated with the initial data in each set of initial modal data in the N sets of initial modal data to obtain N sets of feature vectors; a third determination module, used to determine the learning weight of the initial features of each initial data in the M initial data to obtain M learning weights; and a first calculation module, used to perform weighted calculation on a set of feature vectors associated with a set of initial modal data and the learning weights of the initial features associated with a set of initial modal data for a set of initial modal data to obtain a target fusion data.

[0084] Optionally, in the regional risk determination device provided in this application embodiment, the second acquisition unit 51 includes: a second classification module, used to perform modal classification on M initial data according to the scene characteristics of each initial data to obtain N sets of initial modal data, wherein each set of initial modal data is associated with a scene type; a first acquisition module, used to acquire Y scene categories, determine the predicted probability that each set of initial modal data belongs to Y scene categories, and obtain N sets of predicted probability sets, wherein each set of predicted probability sets includes Y predicted probabilities, where Y is a positive integer; and a fourth determination module, used to determine the prediction weight of each set of predicted probability sets to obtain N prediction weights, and perform weighted calculation on each prediction weight and each set of predicted probability sets to obtain N target fusion data.

[0085] Optionally, in the regional risk determination device provided in this application embodiment, the second acquisition unit 51 includes: a second conversion module, used to perform vector conversion on the initial features of M initial data to obtain M feature vectors; a second calculation module, used to calculate the normalization factors of the feature vector with M-1 feature vectors respectively using a probability allocation function to obtain M-1 normalization factors, and use the M-1 normalization factors to determine the overlap data of the feature vector; a fifth determination module, used to determine the overlap data of the feature vector with M-1 feature vectors respectively for a feature vector to obtain M-1 overlap data; and a construction module, used to construct a feature matrix based on the M overlap data and the M overlap data, and determine N target fusion data based on the feature matrix.

[0086] Optionally, in the regional risk determination device provided in the embodiments of this application, the input unit 52 includes: an extraction module, used to extract risk features from the risk prediction results, wherein the risk features are used to characterize the degree of correlation of each target fusion data with P risk categories, where P is a positive integer; and a sixth determination module, used to determine the level of the P risk categories according to the risk features, obtain P risk levels, and generate an early warning strategy from the P risk levels.

[0087] Optionally, in the regional risk determination device provided in this application embodiment, the input unit 52 includes: a second acquisition module, used to acquire Q historical data over a historical time period and acquire an initial risk prediction model, where Q is a positive integer; a training module, used to perform unsupervised training on the initial risk prediction model using the Q historical data to obtain a first prediction model; a third acquisition module, used to acquire P risk categories, use the P risk categories to label the Q historical data to obtain Q historical category data, and use the Q historical feature data to perform supervised training on the first prediction model to obtain a second prediction model; and a processing module, used to perform knowledge distillation processing on the second prediction model to obtain a risk prediction model.

[0088] The aforementioned risk determination device includes a processor and a memory. The first acquisition unit 50, the second acquisition unit 51, the input unit 52, etc., are all stored in the memory as program units. The processor executes the aforementioned program units stored in the memory to realize the corresponding functions.

[0089] The processor contains a kernel, which retrieves the corresponding program units from memory. One or more kernels can be configured, and adjusting kernel parameters can address the issues of insufficient accuracy and timeliness in identifying and issuing early warnings of regional risks in related technologies.

[0090] The memory may include non-permanent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0091] This invention provides a computer storage medium for storing a program, wherein the program, when running, controls the device where the computer storage medium is located to execute a method for determining regional risk.

[0092] Figure 6 This is a schematic diagram of an electronic device provided according to an embodiment of this application, such as... Figure 6 As shown, this embodiment of the invention provides an electronic device 60, which includes a processor, a memory, and a program stored in the memory and executable on the processor. The processor is used to execute computer-readable instructions, wherein the computer-readable instructions, when executed, perform a method for determining regional risk. The device described herein may be a server, PC, PAD, mobile phone, etc.

[0093] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of a method for determining regional risk according to various embodiments of this application.

[0094] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0095] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0096] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0097] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0098] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0099] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0100] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0101] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0102] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for determining regional risk, characterized in that, include: Obtaining risk prediction requirements for a target area, and determining target data types based on the risk prediction requirements, includes: identifying the risk prediction requirements to determine environmental and infrastructure information for the target area; determining the target data types required for predicting the risks of the target area based on the environmental and infrastructure information; wherein, the target data types are used to characterize the data types required for predicting the risks of the target area; Initial data conforming to the target data type are obtained from different data sources to obtain M initial data, and the M initial data are fused to obtain N target fused data, where M is greater than N, and M and N are positive integers; The fused data of the N targets are input into the risk prediction model, and the risk prediction result is output. The risk prediction result is the probability distribution of one or more risk events. After obtaining the risk prediction result, an early warning strategy is determined based on the risk prediction result.

2. The method according to claim 1, characterized in that, Initial data conforming to the target data type is obtained from different data sources, resulting in M ​​initial data sets, including: Multiple data sources are determined based on the target data type, wherein each data source refers to a device in the target region that collects the M initial data. Based on the risk prediction requirements, the screening criteria are determined, and K data points to be processed are extracted from the data source according to the screening criteria. The K data points to be processed are then preprocessed to obtain the M initial data points, where K is greater than or equal to M and K is a positive integer.

3. The method according to claim 1, characterized in that, The M initial data points are fused to obtain N target fused data points, including: Based on the scene characteristics of each initial data point, the M initial data points are classified into modes to obtain N sets of initial mode data, where each set of initial mode data is associated with a scene type; The initial features associated with the initial data in each of the N initial modal data sets are transformed into vectors to obtain N sets of feature vectors. The learning weights of the initial features of each of the M initial data are determined to obtain M learning weights; For a set of initial modal data, a set of feature vectors associated with the set of initial modal data and the learning weights of the initial features associated with the set of initial modal data are weighted and calculated to obtain a target fusion data.

4. The method according to claim 1, characterized in that, The M initial data points are fused to obtain N target fused data points, including: Based on the scene characteristics of each initial data point, the M initial data points are classified into modes to obtain N sets of initial mode data, where each set of initial mode data is associated with a scene type; Obtain Y scene categories, determine the predicted probability of each set of initial modal data belonging to the Y scene categories, and obtain N sets of predicted probabilities, where each set of predicted probabilities includes Y predicted probabilities, where Y is a positive integer; Determine the prediction weights for each set of prediction probabilities to obtain N prediction weights. Perform a weighted calculation on each prediction weight and each set of prediction probabilities to obtain the N target fusion data.

5. The method according to claim 1, characterized in that, The M initial data points are fused to obtain N target fused data points, including: The initial features of the M initial data are transformed into M feature vectors; For a feature vector, the normalization factors of the feature vector and M-1 feature vectors are calculated using the probability allocation function to obtain M-1 normalization factors, and the reciprocity data of the feature vector is determined using the M-1 normalization factors. For a feature vector, determine the degree of overlap between the feature vector and M-1 feature vectors respectively, and obtain M-1 degree of overlap data; A feature matrix is ​​constructed based on M overlap data and M counterbalancing data, and the N target fusion data are determined based on the feature matrix.

6. The method according to claim 1, characterized in that, The early warning strategy is determined in the following way: Risk features are extracted from the risk prediction results, wherein the risk features are used to characterize the degree of association between each target fusion data and P risk categories, where P is a positive integer; Based on the risk characteristics, the levels of the P risk categories are determined to obtain P risk levels, and the early warning strategy is generated from the P risk levels.

7. The method according to claim 1, characterized in that, The risk prediction model is trained in the following way: Obtain Q historical data points for a given historical time period and develop an initial risk prediction model, where Q is a positive integer; The initial risk prediction model is trained unsupervised using the Q historical data to obtain the first prediction model; Obtain P risk categories, use the P risk categories to label the Q historical data to obtain Q historical category data, and use the Q historical feature data to perform supervised training on the first prediction model to obtain the second prediction model; The risk prediction model is obtained by performing knowledge distillation on the second prediction model.

8. A device for determining regional risk, characterized in that, include: The first acquisition unit is used to acquire risk prediction requirements for a target area and determine target data types based on the risk prediction requirements, including: identifying the risk prediction requirements to determine environmental information and infrastructure information of the target area; and determining the target data type required for predicting the risks of the target area based on the environmental information and the infrastructure information; wherein the target data type is used to characterize the data type required for predicting the risks of the target area. The second acquisition unit is used to acquire initial data that conforms to the target data type from different data sources, obtain M initial data, and fuse the M initial data to obtain N target fused data, where M is greater than N and M and N are positive integers; The input unit is used to input the fused data of the N targets into the risk prediction model and output the risk prediction result, wherein the risk prediction result is the probability distribution of one or more risk events. After obtaining the risk prediction result, an early warning strategy is determined based on the risk prediction result.

9. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method for determining regional risk as described in any one of claims 1 to 7.

10. An electronic device, characterized in that, It includes one or more processors and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to implement the method for determining regional risk as described in any one of claims 1 to 7.

Images