A city security risk grading assessment early warning system

Abstract

The present application relates to the technical field of security risk classification, in particular to a kind of city security risk classification and evaluation early warning system, comprising: risk sampling module, region division module, risk load module, risk progression module and risk update module;Through the risk data and risk type in the risk sampling area are listed;Using the risk type and the risk load capacity of each sampling area is classified, determine the regional risk distribution map of each sampling area;From the regional risk distribution map, extract risk point, with the data proportion of each risk point and the risk load capacity of region, generate regional risk load probability;Set the risk point load capacity of each risk point, and extract the progression probability of risk point load capacity, obtain the relative risk state of each sampling area;According to the relative risk state of each sampling area, identify the risk entity and the update range of risk entity of each sampling area, and iteratively update the risk level of each sampling area;The efficiency of risk level division is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of security risk classification, in particular to a city security risk classification assessment and early warning system. BACKGROUND

[0002] In modern society, city security risk assessment is a particularly important issue. As an important place for human life, work and entertainment, the safety of the city directly affects the quality of life and social stability of people. When classifying city security risks, a static scorecard model is generally used, which cannot reflect the dynamic evolution characteristics of risks. At the same time, since spatial analysis is based on the coarse-grained division of administrative boundaries, it is difficult to locate the risk hotspots that occur when classifying risks, resulting in problems such as delay and reduced accuracy in locating city risks.

[0003] For example, Chinese patent publication CN118297409A discloses a city underground security risk assessment method, device, electronic equipment and storage medium. It belongs to the technical field of geological disaster assessment. The method includes: obtaining risk factors affecting city underground security; normalizing the risk factors according to a pre-constructed risk assessment matrix to obtain normalized risk factors; constructing a city underground security risk assessment model structure based on the risk factors and the causal relationships between the risk factors; determining a city underground security risk assessment model according to the normalized risk factors and the city underground security risk assessment model structure; obtaining security risk data of the city underground to be evaluated, and inputting the security risk data into the city underground security risk assessment model to obtain an evaluation result of the city underground to be evaluated.

[0004] For example, Chinese patent publication CN117952413A discloses a city security risk assessment method and system, relating to the technical field of security risk assessment, which includes collecting city security risk information in a certain area through city Internet of Things devices; the city security risk assessment center obtains city security risk information collected by the Internet of Things devices, and compares and identifies city security risk categories with historical data in the database; after confirming the city security risk categories, the risk source detailed information is collected through the Internet of Things devices to evaluate the city security risk level and send a notification. The present application has the beneficial effect that it uses Internet of Things devices to monitor the city environment in real time, collects various types of security risk information, compares and analyzes the historical data in the database, identifies the risk types and evaluates the risk level.

[0005] Existing technologies use Bayesian networks to describe accident types and risk sources by describing their status. However, these technologies only consider risk sources in a single area and do not comprehensively classify the distribution of multiple risk points in a city. This results in insufficient data samples and inadequate identification of risk area correlations during risk classification. Consequently, the comprehensive classification between risk areas cannot be quickly identified through the focal areas of each risk source, which is not conducive to overall decision-making and observation of urban risks. Summary of the Invention

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: an urban safety risk classification assessment and early warning system, comprising: a risk sampling module, used to sample urban safety risks, list risk data within the sampling area, and describe the risk type within the sampling area using the risk data.

[0007] The region division module is used to classify regions based on their risk type and risk load capacity, and to determine the regional risk distribution map for each sampling region.

[0008] The risk load module is used to extract risk points from the regional risk distribution map, and to quantitatively assess the regional risk load capacity of each sampled area based on the data proportion of each risk point in the regional risk distribution map, thereby generating the regional risk load probability.

[0009] The risk progression module is used to identify the risk radius of each risk point within the sampling area, combine the risk type within the risk radius with the regional risk load probability, set the risk point load capacity of each risk point, and extract the progressive probability of the risk point load capacity to obtain the relative risk status of each sampling area.

[0010] The risk update module is used to construct a risk event network analysis based on the relative risk status of each sampling area, identify the risk entities and update range of each risk entity under each sampling area, and perform iterative updates of the risk level of each sampling area in combination with the risk evolution path of the risk event network.

[0011] The beneficial effects of this invention are as follows: First, this invention uses a data structure of sampling points, risk types, and feature gradients to sample the distribution of urban safety risk data in a grid form, and sets risk labels for each risk type to construct data content about the distribution of risk types in the city. The probability of each risk type can be dynamically displayed, thereby improving the accuracy of risk assessment and setting.

[0012] Second, this invention optimizes the efficiency of regional risk classification by adjusting the risk probability of each region using relative data from multiple time intervals, classifying the risk types in each region, and introducing the regional risk load capacity set relative to each region. By using the probability interval after risk probability adjustment as the basis for setting the classification region, and introducing the number of classifications and the area of ​​the region after classification, the number of classifications is adjusted to achieve an adaptive adjustment of regions when dividing risk regions, which improves the cycle of regional classification and shortens the update cycle of urban risk data when risks occur.

[0013] Third, this invention extracts labels from a regional risk distribution map based on the proportion of risk types, regional risk load capacity, and regional risk load probability, sets combined labels for each risk point, and describes the regional risk load probability of each risk point. Then, it introduces Bayes' theorem and risk radius to dynamically update the progressive probability of each risk point, realizing the mapping relationship between multiple risk points and urban infrastructure carrying capacity. By calculating the corresponding transmission form of its progressive probability, it can identify the corresponding range and update range of each risk entity when the risk occurs, thereby realizing the calibration and iterative update process of dynamic risk transfer and establishing an early warning and protection mechanism related to risk entities, risk evolution paths, and urban infrastructure. Attached Figure Description

[0014] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0015] Figure 1 This is a schematic diagram of an urban safety risk classification assessment and early warning system.

[0016] Figure 2 This is a flowchart illustrating the risk sampling module of an urban safety risk grading assessment and early warning system.

[0017] Figure 3 This is a flowchart illustrating the regional division module of an urban safety risk grading assessment and early warning system.

[0018] Figure 4 This is a flowchart illustrating the risk load module of an urban safety risk classification assessment and early warning system.

[0019] Figure 5 This is a flowchart illustrating the risk progression module of an urban safety risk grading assessment and early warning system.

[0020] Figure 6 This is a flowchart illustrating the risk update module of an urban safety risk classification assessment and early warning system. Detailed Implementation

[0021] The embodiments of the present invention are described in detail below. The embodiments described below are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention. Where specific techniques or conditions are not specified in the embodiments, they shall be performed in accordance with the techniques or conditions described in the literature in the art or in accordance with the product manual.

[0022] See Figure 1 A city safety risk classification assessment and early warning system includes: a risk sampling module, a region division module, a risk load module, a risk progression module, and a risk update module; wherein, the output end of the risk sampling module is connected to the region division module, the output end of the region division module is connected to the risk load module, the output end of the risk load module is connected to the risk progression module, and the output end of the risk progression module is connected to the risk update module.

[0023] The risk sampling module is used to sample urban safety risks, list risk data within the sampling area, and describe the risk types within the sampling area using the risk data.

[0024] The region division module is used to classify regions based on their risk type and risk load capacity, and to determine the regional risk distribution map for each sampling region.

[0025] The risk load module is used to extract risk points from the regional risk distribution map, and to quantitatively assess the regional risk load capacity of each sampled area based on the data proportion of each risk point in the regional risk distribution map, thereby generating the regional risk load probability.

[0026] The risk progression module is used to identify the risk radius of each risk point within the sampling area, combine the risk type within the risk radius with the regional risk load probability, set the risk point load capacity of each risk point, and extract the progressive probability of the risk point load capacity to obtain the relative risk status of each sampling area.

[0027] The risk update module is used to construct a risk event network analysis based on the relative risk status of each sampling area, identify the risk entities and update range of each risk entity under each sampling area, and perform iterative updates of the risk level of each sampling area in combination with the risk evolution path of the risk event network.

[0028] Preferably, when sampling data on urban safety risks, potential risks that may occur in the city, such as natural disasters and man-made accidents, are described, and the form and content of the risk are defined. After defining the risk, data is extracted from the database of urban risks, and the types of risks present in the data are described. The frequency of occurrence, scope of impact, possible losses, and location of each type of risk in the sampling area are marked in the risk data, and the label of the risk type is explained.

[0029] like Figure 2 As shown, the implementation of the risk sampling module also includes: specifying each sampling area, obtaining the sampling points of risk data in each sampling area, and listing the risk data in the sampling area based on the risk types of multiple sampling points.

[0030] Based on the gradient of each risk type, the frequency of occurrence, scope of impact, losses caused, and geographical location of each risk type in the sampling area are characterized, and risk labels corresponding to each risk type are set.

[0031] Determine whether the risk label matches the current sampling area. If it does, calculate the probability of the current risk label appearing at each sampling point based on the risk type. If not, mark the risk label as a risk-free event.

[0032] After adding the risk label to the risk type, the corresponding risk type will be output.

[0033] Preferably, the sampling area includes multiple urban zones such as commercial areas, industrial areas, and residential areas. Then, in each sampling area, historical data is used to represent the problems existing in that area as risk points, such as restaurant kitchen fire risk points in commercial areas and hazardous chemical warehouse leakage risk points in industrial areas. After listing all possible risk locations in each sampling area, the risk points are categorized by risk type, such as fire, theft, environmental pollution, and heavy snowfall, representing urban risk types. Then, for each risk type, the frequency of occurrence, scope of impact, losses caused, and geographical location are described, such as the number of historical occurrences and time periods, for example, an average of 3 fires per year in a certain area. After defining the scope of the risk's impact, the losses caused by the corresponding risk are examined, such as an average loss of 500,000 yuan per fire, to illustrate the specific situation of the corresponding risk type. The geographical location mainly describes the relative coordinates of the risk points, used for subsequent descriptions of the relative impact on the region and individual risk points. Finally, it is determined whether the risk labels describing the risk type can be matched with the location of the sampled area. For example, a residential area is labeled with a red label for hazardous chemical leaks. If there is a clear mismatch, the mismatched risk label is marked as an invalid risk. This completes the risk listing of the current sampled data. These risk types are used to list the risk situations in different areas, serving as the main conditions and main data framework for assessing urban risks.

[0034] Preferably, the gradient of risk type represents the values ​​of frequency of occurrence, scope of impact, loss caused, and geographical location.

[0035] Preferably, determining whether a risk label matches the current sampling area can be done by using preset association rules in historical data to describe the content described by the risk label for the region represented by each sampling area. This helps determine if the risk label for that risk category is correct. If correct, it facilitates the subsequent use of risk types to describe the risk load capacity of the current sampling area. The preset association rules mainly indicate whether the risk type is data that normally appears in the sampling area, preventing data acquisition errors.

[0036] In one embodiment of the present invention, the regional division module is mainly designed to classify and plan each region according to its regional risk load capacity. Regional risk load capacity refers to the comprehensive ability of a specific region to withstand, respond to, and recover from risk events within a unit of time. Regional risk load capacity can be set through the following aspects, such as response capacity, withstand capacity, and recovery capacity. For example, withstand capacity can be described by the disaster resistance capacity of infrastructure, evacuation capacity, and redundancy of key resources, such as the coverage rate of dual power / communication circuits. Response capacity can be described by the speed and completeness of each region's response to risks, such as the timeliness of emergency response, emergency resource reserves, and the completeness of emergency plans. Recovery capacity can be described by selecting dimensions such as economic resilience, social resilience, and ecological restoration capacity to illustrate the speed and ability to recover after a risk occurs.

[0037] At this point, the response capacity, tolerance capacity, and recovery capacity of the current sampling area are represented digitally. The regional risk load capacity of the current sampling area is represented by the conditional probability of these three capabilities and various risk types occurring together. The conditional probability reflects the non-linear relationship between capacity and risk, describing the differentiated grading of regional capacity under different risks. For example, if the current risk type is the leakage of a certain item, then the response capacity, tolerance capacity, and recovery capacity of the facilities in the area are selected with scores of 0.6, 0.75, and 0.5 respectively. This calculates the probability of the leakage risk type occurring when the corresponding capacity is at the corresponding value, thus describing the value of regional risk load capacity in different sampling areas. That is, the index value of regional risk load capacity is based on the conditional probability of the region's preset response capacity, tolerance capacity, and recovery capacity under different risk types.

[0038] Preferred, such as Figure 3 As shown, the implementation of the region division module includes: using the regional risk load capacity of each sampling region to determine whether the basic risk probability when the current risk type occurs is greater than the index value of the regional risk load capacity; if it is greater, the difference between the basic risk probability when the current risk type occurred last time and the index value of the regional risk load capacity is used as the basic risk probability of the current risk type.

[0039] If it is less than the limit, then the risk excess probability is set to the upper limit of the difference between the basic risk probability in the current risk type and the index value of the current regional risk load capacity.

[0040] Based on the risk excess probability identification, the increasing probability difference between the current risk type and the regional risk load capacity is identified in multiple time intervals. The increasing probability difference is used as the probability interval for regional classification. Based on the probability interval for regional classification, a regional risk distribution map is output.

[0041] Preferably, the aforementioned basic risk probability represents the probability value of a certain risk type occurring. This basic risk probability is determined by dividing the number of times the corresponding risk type occurs in historical data by the total number of occurrences of all risk types. Then, this probability value is compared with the regional risk load capacity index value to prevent an excessively high probability of a single risk, which could trigger an overreaction and cause significant data fluctuations when identifying historical risk events in the sampling area. Simultaneously, the regional risk load capacity index value is represented by the basic risk probability of that risk type and the conditional probability corresponding to the facilities' response capacity, resilience, and recovery capacity in the current region. This conditional probability is calculated based on the sampled historical data. The basic risk probability of each type is also calculated based on the sampled historical data. After the basic risk probability is reduced through dynamic adjustment, the basic risk probability of all risk types is identified, and the difference between the basic risk probability and the regional risk load capacity is calculated. The upper limit of this difference is expressed as the absolute value of the difference between the two. Then, according to the time series, the probability difference of the risk excess probability of the difference type at different time points is found to determine whether the risk type has increased or decreased over time, indicating whether the region needs to be adjusted, representing the trend of different risk types over time, and dividing these regions into regional risk distribution maps on the sampled area through the intervals shown by the increasing probability difference.

[0042] That is, when outputting the regional risk distribution map based on the probability intervals of the regional classification, a mapping relationship is constructed between the probability intervals of the regional classification and the sampling area. The area corresponding to the value of each probability interval is regarded as the classification area, and the positions of each classification area are combined to form the regional risk distribution map.

[0043] Preferably, the increasing probability difference is represented by the difference between the risk excess probability of the corresponding risk type in the current risk data in multiple time intervals. This increasing probability difference is used to identify whether the risk in the current area is gradually increasing or decreasing, so as to determine the regional risk and dynamic probability of each area under the risk classification.

[0044] Preferably, the method of determining the regional risk distribution map of each sampling area by classifying the region according to the risk type and regional risk load capacity of each sampling area also includes: loading the mapping relationship between risk type and sampling area, and calculating the number of levels and the size of the leveled area when classifying the region.

[0045] Based on the data distribution of each hierarchical region during regional classification, data clustering is performed using the distance between the center points of each hierarchical region.

[0046] Record the difference in the number of levels in the corresponding hierarchical regions before and after clustering, and output this difference as a note label for the regional risk distribution map.

[0047] Preferably, the number of levels in the regional classification is used to represent the number of levels into which the sampling area is divided. The number of levels is set based on the value of the probability interval. For example, when each risk type is judged by increasing the risk excess probability in the current sampling area, it will form a region about the change of probability value. For example, under four consecutive time intervals, the difference between the values ​​of risk excess probability in two consecutive time intervals forms an interval, such as a probability interval in the form of {+0.1, -0.1, +0.2}. After obtaining the probability interval corresponding to each risk type, multiple sets of values ​​with different change forms can be obtained. Then, multiple risk points representing different regions can be obtained. Then, these points are clustered to obtain a comprehensive regional risk distribution map.

[0048] Preferably, the mapping relationship between the loaded risk type and the sampling area is used to view the risk type corresponding to the graded area in the regional risk distribution map, and how the risk type is recorded in the current sampling area. Then, the number and size of the graded areas are identified to determine the distribution pattern of different risk types in the current sampling area. Next, clustering is performed on the distribution of data corresponding to each graded area and the center point of the corresponding area to identify risk propagation patterns or hotspots in different probability intervals. When generating the final regional risk distribution map, its XY axes are used as the geographical coordinates of the corresponding sampling area. The Y-axis displays the increasing probability difference between the risk type and the regional risk load capacity of each area in multiple time intervals, representing the risk situation generated at the corresponding geographical coordinates.

[0049] Preferably, when calculating the number of levels in the regional classification, the data for each classification region is loaded to obtain the area occupied by each classification region in the sampling region. The size of the classification region is then set as the starting point, with the classification region occupying the largest area as the starting point. Obtain the secondary region of the hierarchical region, and set the secondary region according to a preset ratio. Configure the settings so that the area of ​​the secondary region of the current hierarchical region is located at... arrive Between these points, the secondary regions are continuously divided according to a preset ratio, and the values ​​are calculated separately. , The process continues until no corresponding hierarchical region exists within the divided intervals. The number of divisions is considered the number of levels for the current region. The preset ratio value is a positive integer greater than 0, used to adjust the number of levels in the current hierarchical region division. This value can be set based on the preset ratio value that has been used most frequently in historical data for hierarchical region division, or by selecting the average value of the preset ratio value used in historical data.

[0050] The subsequent data clustering using the distance between the center points of each hierarchical region also includes clustering based on the nearest hierarchical region to the center point of any hierarchical region. This nearest hierarchical region represents the region with the smallest distance between the center points of two adjacent hierarchical regions. This clustering is repeated until the average distance between the center points of each hierarchical region in the multiple regions formed by each hierarchical region is equal, resulting in multiple regions after hierarchical region clustering. The hierarchical number of the clustered regions is calculated in the same way as the hierarchical region. Then, the difference between the hierarchical numbers before and after clustering is calculated as the output annotation label. The multiple regions after clustering are used as the output regional risk distribution map.

[0051] In one embodiment of the present invention, the risk load module is mainly used to identify the location of risk points in the regional risk distribution map and set the regional risk load probability that the risk points can correspond to.

[0052] Preferably, when extracting risk points, risk points are selected from the regions after hierarchical regional clustering to represent the proportion of data representing regional risks in multiple regions after clustering to the total sampled data. Subsequently, when describing the regional risk load capacity of each sampled region, it can be understood as integrating the conditional probabilities of the regional risk load capacity corresponding to the current risk point's response capacity, tolerance capacity, and recovery capacity for the corresponding risk type to obtain the regional risk load probability.

[0053] For example, the ratio of data sampled for a single risk type corresponding to a risk point within the current sampling area is set as the type proportion. This type proportion is multiplied by the regional risk load capacity index value. The average of this product is then calculated based on the number of risk types that the current risk point's region can correspond to. This average value is used as the regional risk load probability. Finally, the regional risk load probability for multiple risk points is calculated for each risk point's region.

[0054] like Figure 4As shown, the implementation of the risk load module also includes: extracting labels from the regional risk distribution map and setting combined labels for each risk point; combining the risk types and remarks labels existing in each cluster area of ​​the regional risk distribution map into combined labels for the current risk point.

[0055] The risk type of the corresponding risk point is determined from the group label. Based on the risk type of each risk point, the proportion of the corresponding type is determined. Based on the proportion of the type and the regional risk load capacity, the regional risk load probability of the corresponding risk point is described.

[0056] In one embodiment of the present invention, the risk progression module mainly describes the risk type corresponding to the risk radius of the risk point to identify the comprehensive data situation of the corresponding risk point.

[0057] Preferably, the risk radius of each risk point is based on the radius of the smallest inscribed circle of multiple regions in the regional risk distribution map generated by the final clustering in the regional segmentation module. This risk radius is used to identify urban safety risks occurring in different regions. When the risk radii of different regions are accurately set, it is easier to identify risk boundary areas, promoting cooperation in risk management and emergency response between different regions. This helps to form a unified risk management strategy and improve the overall risk response capability.

[0058] Preferably, the risk load capacity of a risk point is biased towards describing the number of risk events occurring in the area corresponding to the current risk point. When the number of risk events changes, the risk load capacity of each area is affected under different classification conditions, and the impact is regarded as the risk point load capacity.

[0059] The progressive probability indicates whether each risk point exhibits linear or exponential decay within its corresponding risk radius, and uses the decay pattern as the relative risk state of each sampling area.

[0060] like Figure 5 As shown, the implementation of the risk progression module includes: setting the risk radius of each risk point using the smallest inscribed circle of multiple regions in the regional risk distribution map; updating the regional risk load probability of each risk point using Bayes' theorem based on the risk type and corresponding regional risk load probability within the risk radius; setting the risk point load capacity of each risk point, where the load capacity of each risk point is the probability value updated using Bayes' theorem; then determining whether the risk radii of each risk point overlap; if they overlap, identifying the risk points corresponding to the overlapping risk radii and calculating the difference in risk point load capacity between the corresponding risk points as the progression probability of the current risk point; and using the decay coefficient of the progression probability as the output of the relative risk state of each sampling region.

[0061] If they do not overlap, the difference in risk point load capacity between the two adjacent risk points is used as the progressive probability of the current risk point, and the decay coefficient of the progressive probability is used as the output of the relative risk state of each sampling area.

[0062] Preferably, the method for updating the probability value of each risk point using Bayes' theorem is as follows: Assume that the risk type and risk point are represented by j and i respectively, and there are m risk types, where the number of risk types and risk points is set based on their quantity within the corresponding area of ​​the regional risk distribution map. Then, the regional risk load probability is updated.

[0063] Its formula can be expressed as: ;in, This represents the risk load capacity of risk point i. The risk load capacity is expressed as the updated and adjusted regional risk load probability calculated for risk point i, illustrating the value of a single risk point after multiple updates and iterations. The upper part of this formula describes the relative product of the type proportion and the conditional probability represented by the regional risk load capacity, while the lower part describes the value after multiple iterations when the number of events of risk type j contained in risk point i changes, illustrating whether the probability value will change accordingly in multiple time intervals. This indicates the percentage of risk type j. Indicates risk type j in The index value of regional risk load capacity under the circumstances, The number of events of risk type j within the risk radius of the current risk point i is used to illustrate the relative value of the regional risk load capacity under the data proportion corresponding to the current risk type. The number of events is also linked to the type proportion. When sampling urban safety, each data point represents a risk event. Indicates the number of steps updated. The number of events of risk type j within the risk radius of the current risk point i at the k-th update step. Indicates risk type j in The regional risk load capacity index value, under certain circumstances, primarily indicates whether the regional risk load capacity index value will change due to the amount of data during multiple iterations or when data sampling and iterative updates are performed using multiple time intervals for the same risk point and risk type. For general cases, the regional risk load capacity for risk type j is expressed as... ;in, This represents the base risk probability of risk type j. This represents the probability of risk type j corresponding to the scores of coping ability, resilience, and recovery ability in the current region; subsequently, if the probability value obtained under the relative time quantity is identified, the number of events is labeled to obtain this value.

[0064] If we then denote the overlapping risk points as i1 and i2, then their attenuation coefficients... It can be represented as: ;in, This represents the difference in load capacity between risk points i1 and i2. Represents an exponential constant. This represents the linear decay exponent, typically set to 0.001. This represents the distance between risk points i1 and i2. The output attenuation coefficient will be used as the label corresponding to its sampling area. The progressive probability of overlap will be summed to indicate whether the risk of the corresponding area is high or low. For example, 0.8 and 0.5 can be used as dividing nodes. The part greater than 0.8 is considered high risk, the part between 0.5 and 0.8 is considered medium risk, and the rest is considered low risk.

[0065] If there is no overlap, let the two adjacent risk points of the current risk point i1 be i3 and i4, and their attenuation coefficients be... Then it is represented as ;in, This represents the difference in load capacity between the two adjacent risk points i3 and i4 of the current risk point i1; ,in , and These represent the load capacity of risk points i1, i3, and i4, respectively. This represents the average distance between two adjacent risk points i3 and i4 of the previous risk point i1. The attenuation coefficient obtained at this time takes into account the difference between the two adjacent risk points and the current risk point to illustrate their relative risk situation; the subsequent processing method is the same as that for the attenuation coefficient when there is overlap, mainly to indicate whether there are areas with high urban safety risks.

[0066] In one embodiment of the present invention, such as Figure 6 As shown, the risk update module is implemented by taking the relative risk status as a starting point, combining the risk entities associated with each risk point, and setting up a risk event network.

[0067] Determine whether there is a correlation between risk entities in the risk event network. If a correlation exists, set the update range of the risk entity based on the updated data each time a correlation occurs.

[0068] The shortest connection path for updating the risk entity's scope at each association is set as the risk path. The differential path of the risk path at each update is set as the risk evolution path. The risk level is then iteratively updated using the risk evolution path.

[0069] Preferably, when setting up the risk event network, the relative risk status of each risk point is used as the edge weight, and the risk point load capacity of each risk point is used as its risk point value. The risk points with correlation are connected to form the risk event network.

[0070] In scenarios involving heavy rain, drainage system failure, and subsequent flooding, the flooded areas are treated as individual risk entities. The system then analyzes whether these risk entities are correlated during the flooding event. If correlated, these risk entities are grouped into a risk event network based on their correlation. The system then determines the updated data range for each correlation, such as expanding the flood risk risk from a 3km radius to a 4km radius. Risk entities in this expanded area are marked as the updated range. The system then identifies the shortest connection path for associating risk entities within this updated range and compares the differences between these paths after each risk entity update. This analysis explains how the risk evolves under a risk event, facilitating iterative updates to the regional risk classification. This allows for dynamic capture of risk evolution paths, enabling precise perception and early warning from local risk points to the overall risk situation.

[0071] To assist subsequent staff in making decisions on current urban safety risks based on multiple probability values ​​such as the risk evolution path and the load capacity of the corresponding risk points, thereby improving the setting and correlation of various risks.

[0072] Ultimately, the processing sequence of regional risk sampling and planning → regional load capacity classification and verification → geospatial risk radius analysis → dynamic updating of risk logical relationships is achieved, so as to complete the classification data of urban safety risks in each region in real time.

[0073] Subsequently, when updating the risk level through the risk evolution path, the risk level is updated primarily by the rate of change of the relative values ​​and weights of the risk entities corresponding to the risk evolution path.

[0074] For example, obtain the path weight change rate of each risk entity on the risk evolution path, and update the risk level of each risk entity in combination with the relative risk status before the update.

[0075] At this point, a sensitivity coefficient is introduced. This coefficient can be 0.1 or other values. The updated risk level is obtained by multiplying the sensitivity coefficient by the path weight change rate and then adding the relative risk state before the risk entity update. Alternatively, risk entities on the risk evolution path can be incorporated into the regional risk distribution map to obtain the reclassified hierarchical regions of multiple areas before and after the addition of risk entities. These hierarchical regions are then used as the content for iterative updates.

[0076] The aforementioned path weight change rate represents the rate of change of the sum of the weights of each risk point edge in the risk event network along the corresponding risk evolution path. It describes whether a larger risk range and more severe risk situations will arise as the data is updated.

[0077] Preferably, the logical transmission relationship of each risk point indicates whether there is a correlation between the risk entities represented by each risk point. Only risk entities that are correlated can be combined into a risk event network. At this time, the correlation of risk entities will be pre-set in the database with correlation rules, which will facilitate the subsequent establishment of the parts that need risk warning in each sampling area. These iteratively updated parts will be used for subsequent alarm decisions to complete the graded processing of urban safety risks.

[0078] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention, which are still covered within the protection scope of the present invention.

Claims

1. A city safety risk classification assessment and early warning system, characterized in that, include: The risk sampling module is used to sample urban safety risks, list risk data within the sampling area, and describe the risk types within the sampling area using the risk data. The region division module is used to classify regions based on the risk type and regional risk load capacity of each sampling region, and determine the regional risk distribution map of each sampling region. Regional risk load capacity refers to the comprehensive ability of a specific region to withstand, respond to, and recover from risk events within a unit of time. The response capacity, withstand capacity, and recovery capacity of the current sampled region are represented digitally, and the regional risk load capacity of the current sampled region is represented by the conditional probability of the co-occurrence of these three capabilities and various risk types. The risk load module is used to extract risk points from the regional risk distribution map, and to quantitatively assess the regional risk load capacity of each sampled area based on the data proportion of each risk point in the regional risk distribution map, thereby generating the regional risk load probability. The risk progression module is used to identify the risk radius of each risk point within the sampling area, combine the risk type within the risk radius and the regional risk load probability, set the risk point load capacity of each risk point, and extract the progressive probability of the risk point load capacity to obtain the relative risk status of each sampling area. The risk update module is used to construct a risk event network analysis based on the relative risk status of each sampling area, identify the risk entities and update range of each risk entity under each sampling area, and iteratively update the risk level of each sampling area in combination with the risk evolution path of the risk event network. Starting from the relative risk status, the risk entities associated with each risk point are combined to set up a risk event network; The implementation method for determining the regional risk distribution map of each sampling area based on the risk type and regional risk load capacity of each sampling area also includes: loading the mapping relationship between risk type and sampling area, calculating the number of levels and the size of the level area when classifying the area; clustering the data using the distance between the center points of each level area based on the data distribution of each level area when classifying the area; recording the difference in the number of levels of the corresponding level area before and after clustering, and outputting the difference in the number of levels as the annotation label of its regional risk distribution map. Data clustering using the distance between the center points of each graded region also includes: clustering by the nearest graded region to the center point of any graded region until the average distance between the center points of each graded region in the multiple regions composed of each graded region is equal, and using the multiple regions after clustering as the output regional risk distribution map. The implementation of the risk load module also includes: extracting labels from the regional risk distribution map and setting combined labels for each risk point; combining the risk types and remarks labels existing in each cluster region of the regional risk distribution map into combined labels for the current risk point; determining the risk type of the corresponding risk point from the group labels; determining the type proportion of each risk point based on the risk type of each risk point; and describing the regional risk load probability of the corresponding risk point based on the type proportion and the regional risk load capacity. The ratio of the data sampled for a single risk type corresponding to a risk point in the current sampling area is set as the type proportion. The type proportion is multiplied by the index value of the regional risk load capacity. The average value of the product is calculated according to the number of risk types that the current risk point can correspond to. This average value is used as the regional risk load probability. Finally, the regional risk load probability of multiple risk points is calculated for the region where each risk point is located. The risk progression module is implemented as follows: It sets the risk radius of each risk point using the smallest inscribed circle of multiple regions in the regional risk distribution map; updates the regional risk load probability of each risk point using Bayes' theorem based on the risk type within the risk radius and the corresponding regional risk load probability; sets the risk point load capacity of each risk point; determines whether the risk radii of each risk point overlap; if they overlap, it identifies the risk points corresponding to the overlapping risk radii and calculates the difference in risk point load capacity between the corresponding risk points as the progression probability of the current risk point; and uses the decay coefficient of the progression probability as the output of the relative risk state of each sampling region. If they do not overlap, it uses the difference in risk point load capacity between the current risk point and two adjacent risk points as the progression probability of the current risk point, and uses the decay coefficient of the progression probability as the output of the relative risk state of each sampling region.

2. The urban safety risk classification assessment and early warning system according to claim 1, characterized in that, The risk sampling module can also be implemented in the following ways: Specify each sampling area, obtain the sampling points of risk data in each sampling area, and list the risk data in the sampling area according to the risk type of multiple sampling points; Based on the gradient of each risk type, the frequency of occurrence, scope of impact, losses caused, and geographical location of each risk type in the sampling area are characterized, and risk labels corresponding to each risk type are set. Determine if the risk label matches the current sampling area. If it does, calculate the probability of the current risk label appearing at each sampling point based on the risk type. If not, mark the risk label as a risk-free event. After adding the risk label to the risk type, the corresponding risk type will be output.

3. The urban safety risk classification assessment and early warning system according to claim 2, characterized in that, The implementation methods of the region division module include: By utilizing the regional risk load capacity of each sampling area, it is determined whether the basic risk probability when the current risk type occurs is greater than the index value of the regional risk load capacity. If it is greater, the difference between the basic risk probability when the current risk type occurred last time and the index value of the regional risk load capacity is used as the basic risk probability of the current risk type. If it is less than, then the risk excess probability is set to the upper limit of the difference between the basic risk probability in the current risk type and the index value of the current regional risk load capacity. Based on the risk excess probability identification, the increasing probability difference between the current risk type and the regional risk load capacity is identified in multiple time intervals, and the increasing probability difference is used as the probability interval for regional classification. A mapping relationship is established between the probability intervals of the regional classification and the sampling area. The area corresponding to the value of each probability interval is regarded as the classification area, and the positions of each classification area are combined to form a regional risk distribution map.

4. The urban safety risk classification assessment and early warning system according to claim 3, characterized in that, The number of levels when calculating the regional classification also includes: Load the data for each graded region and obtain the area occupied by each graded region in the sampling area. Starting from the graded region with the largest area, obtain the secondary region of that graded region and set the secondary region according to a preset ratio. Continue to divide the secondary region according to the preset ratio until there is no corresponding graded region in the divided interval. The number of divisions is regarded as the grade number when the current region is graded.

5. The urban safety risk classification assessment and early warning system according to claim 1, characterized in that, The risk update module can be implemented in the following ways: Determine whether there is a correlation between the risk entities in the risk event network. If a correlation exists, set the update range of the risk entity based on the updated data at each correlation. The shortest connection path for updating the risk entity's scope at each association is set as the risk path. The differential path of the risk path at each update is set as the risk evolution path. The risk level is then iteratively updated using the risk evolution path.

6. The urban safety risk classification assessment and early warning system according to claim 5, characterized in that, The methods for implementing iterative updates of risk levels based on risk evolution paths include: Obtain the path weight change rate of each risk entity on the risk evolution path, and update the risk level of each risk entity in combination with the relative risk status before the update.

Images