Disaster-causing similar typhoon retrieval model acquisition method, retrieval method, terminal and medium

Abstract

The application provides a disaster-causing similar typhoon retrieval model acquisition method, a retrieval method, a terminal and a medium. The method comprises the following steps: acquiring spatial distribution information and meteorological field information of each path observation point in a target typhoon and related typhoons, and dividing the target typhoon into a first target typhoon and a second target typhoon, and dividing each related typhoon into a first related typhoon and a second related typhoon; based on the path similarity and the meteorological field similarity between the first target typhoon and the corresponding first related typhoon, and in combination with the fitting similarity between the second target typhoon and the corresponding second related typhoon, the weight parameters in the disaster-causing similar typhoon search model are optimized and configured to obtain a final disaster-causing similar typhoon search model; and the configured disaster-causing similar typhoon search model is used to obtain the corresponding disaster-causing similar typhoon of a typhoon to be retrieved. The application improves the prediction ability of the disaster-causing similar typhoon for the path abnormal typhoon and improves the accuracy of the disaster-causing similar typhoon search.

Description

Technical Field

[0001] This invention relates to the field of meteorological disaster risk assessment technology, and in particular to a method for obtaining disaster-causing typhoon similarity retrieval models, a retrieval method, a terminal, and a computer storage medium. Background Technology

[0002] Typhoons with similar destructive power are those whose damage is similar to that of typhoons. According to regional catastrophology theory, the severity of a typhoon's damage depends primarily on factors such as the hazard of the causative agent, the vulnerability of the affected area, the stability of the disaster-prone environment, and disaster prevention and mitigation capabilities. Among these, the vulnerability of the affected area, the stability of the disaster-prone environment, and disaster prevention and mitigation capabilities are closely related to the underlying surface conditions along the typhoon's path. Therefore, the similarity in the destructive power of typhoons is not only closely related to their destructive capacity but also to their spatial distribution.

[0003] To prevent typhoon disasters and reduce the impact of destructive typhoons, the current main method for typhoon disaster prevention is the search for similar destructive typhoons. Specifically, in the pre-disaster and during-disaster stages, by searching and obtaining historical typhoon cases similar to the target typhoon, research based on similar typhoon cases can enable forecasting and prediction of the target typhoon's path, intensity, and precipitation area, as well as typhoon disaster warnings and risk assessments. In the post-disaster stage, research on similar destructive typhoons can also provide references for disaster diagnosis and analysis, and comprehensive disaster loss assessment. Therefore, the search for similar destructive typhoons plays an important role and has significant meaning for typhoon disaster response.

[0004] However, current methods for searching similar typhoons still primarily rely on path shape similarity. On the one hand, statistical methods or numerical models based solely on path shape similarity have limited predictive ability for anomalous typhoon paths at turning points or inflection points. This means their predictive accuracy for typhoons with anomalous paths at turning points or inflection points is poor, and their robustness to similar typhoon paths outside of immediate timeframes (e.g., beyond 72 hours) is also low. On the other hand, in addition to typhoon paths, typhoon disaster response also requires understanding the intensity and duration of various disaster-causing factors. Existing methods for searching similar typhoons often only consider the similarity in typhoon paths, failing to adequately consider atmospheric circulation field information related to typhoon disasters. This reduces the ability to assess typhoon disasters, resulting in significant differences in the disaster impact of the retrieved similar typhoons compared to the target typhoon, further reducing the accuracy and forecasting performance of the disaster search results. Summary of the Invention

[0005] In view of the shortcomings of the prior art described above, the purpose of this invention is to provide a method for obtaining disaster-causing similar typhoon retrieval models, a retrieval method, a terminal, and a computer storage medium, which can solve the problems of low ability of existing similar typhoon retrieval methods to assess typhoon disasters and poor stability of the accuracy of predicting abnormal typhoon paths.

[0006] To achieve the above and other related objectives, this invention first provides a method for obtaining a disaster-causing typhoon similarity retrieval model, comprising: acquiring spatial distribution information and meteorological field information of observation points along the path of a target typhoon, and acquiring spatial distribution information and meteorological field information of observation points along the path of related typhoons; wherein, the related typhoons are historical typhoons within the observation area associated with the target typhoon; based on a preset path division method, the target typhoon is divided into a first target typhoon and a second target typhoon, and each of the related typhoons is divided into a first related typhoon and a second related typhoon; based on the correspondence between the target typhoon and the related typhoons, each first related typhoon corresponding to the first target typhoon is determined, and each first related typhoon corresponding to the first target typhoon is determined. The second target typhoon corresponds to the second related typhoon; using the line similarity calculation method, the path similarity between the first target typhoon and the corresponding first related typhoon is obtained, and using the field similarity calculation method, the meteorological field similarity between the first target typhoon and the corresponding first related typhoon is obtained; based on the path similarity and meteorological field similarity between the first target typhoon and the corresponding first related typhoon, combined with the fitting similarity between the second target typhoon and the corresponding second related typhoon, the weight parameters in the disaster-causing similar typhoon search model are optimized to obtain the final disaster-causing similar typhoon search model; wherein, the disaster-causing similar typhoon search model is a weighted model based on the path similarity and the meteorological field similarity.

[0007] As a preferred embodiment of the first aspect, before optimizing the weight parameters in the disaster-causing similar typhoon search model, the method further includes: performing standardization processing on the path similarity and meteorological field similarity between the first target typhoon and the corresponding first related typhoon, so as to perform subsequent steps based on the standardized path similarity and meteorological field similarity.

[0008] As a preferred embodiment of the first aspect above, the method for obtaining the disaster-causing similar typhoon retrieval model further includes: selecting the target typhoon from among the historical typhoons in the observation area, and selecting related typhoons corresponding to the target typhoon.

[0009] As a preferred embodiment of the first aspect above, the method for obtaining the relevant typhoons corresponding to the target typhoon includes: extracting the spatial relationship between the target typhoon and other historical typhoons based on the path distribution of the target typhoon and the path distribution of other historical typhoons, and detecting whether the spatial relationship meets the preset spatial conditions. If so, other historical typhoons that meet the spatial relationship are taken as the relevant typhoons of the target typhoon.

[0010] As a preferred embodiment of the first aspect above, the method for dividing the target typhoon into the first target typhoon and the second target typhoon includes: for each path observation point in the target typhoon, classifying the typhoon sub-paths corresponding to the n path observation points located at the end of the entire path as the second target typhoon; classifying the typhoon sub-paths corresponding to the other path observation points in the target typhoon as the first target typhoon; and the method for dividing the target typhoon into the first target typhoon and the second target typhoon includes: for each path observation point in the related typhoon path, classifying the typhoon sub-paths corresponding to the n path observation points located at the end of the entire path as the second related typhoon; classifying the typhoon sub-paths corresponding to the other path observation points in the related typhoon as the first related typhoon; wherein n is any value from 3 to 5.

[0011] As a preferred embodiment of the first aspect, the method for optimizing the weight parameters in the disaster-causing typhoon similarity search model includes: dividing the value range of the weight parameters in the disaster-causing typhoon similarity search model based on a preset weight interval to obtain initial models under different weight configurations; based on the path similarity and meteorological field similarity between the first target typhoon and the corresponding first related typhoon, using each of the initial models, obtaining the total similarity between the first target typhoon and the corresponding first related typhoon under each weight configuration; comparing the magnitude of each total similarity, and determining the optimal first related typhoon corresponding to each first target typhoon under each weight configuration based on the comparison results; based on the correspondence between the first target typhoon and the optimal first related typhoon under each weight configuration, determining the optimal second related typhoon corresponding to each second target typhoon; using a path fitting similarity evaluation method, obtaining the fitting similarity between the second target typhoon and the corresponding optimal second related typhoon under different weight configurations; comparing the magnitude of each fitting similarity, and taking the weight configuration corresponding to the maximum fitting similarity as the optimal weight configuration, thereby obtaining the final disaster-causing typhoon similarity search model.

[0012] As a preferred embodiment of the first aspect above, for a single weight configuration, the method for obtaining the fitting similarity between the second target typhoon and the corresponding optimal second related typhoon includes: obtaining the distance deviation between each second target typhoon and the corresponding optimal second related typhoon, and obtaining the angle deviation between each second target typhoon and the corresponding optimal second related typhoon; based on preset statistical features, synthesizing each distance deviation to obtain a distance deviation feature value; and based on preset statistical features, synthesizing each angle deviation to obtain an angle deviation feature value; and using the average of the distance deviation feature value and the angle deviation feature value as the fitting similarity between the second target typhoon and the corresponding optimal second related typhoon.

[0013] As a preferred embodiment of the first aspect above, the method for obtaining the distance deviation feature value includes: after obtaining the distance deviation between each second target typhoon and the corresponding optimal second related typhoon, extracting the mean, median, and quartile of each distance deviation as the distance deviation feature value between the second target typhoon and the optimal second related typhoon; and the method for obtaining the angle deviation feature value includes: after obtaining the angle deviation between each second target typhoon and the corresponding optimal second related typhoon, extracting the mean, median, and quartile of each angle deviation as the angle deviation feature value between the second target typhoon and the optimal second related typhoon.

[0014] To achieve the above and other related objectives, the present invention also provides a method for retrieving disaster-causing similar typhoons, comprising: selecting a target typhoon within an observation area, and selecting related typhoons corresponding to the target typhoon; acquiring spatial distribution information and meteorological field information of the target typhoon at each path observation point, and acquiring spatial distribution information and meteorological field information of each related typhoon at each path observation point; using a line similarity calculation method to acquire the path similarity between the target typhoon and each related typhoon, and using a field similarity calculation method to acquire the meteorological field similarity between the target typhoon and each related typhoon; and, based on the path similarity and meteorological field similarity between the target typhoon and each related typhoon, using a disaster-causing similar typhoon search model to acquire the total similarity between the target typhoon and each related typhoon, and taking the similar typhoon corresponding to the maximum total similarity as the disaster-causing similar typhoon of the target typhoon; wherein, the disaster-causing similar typhoon search model is a model obtained based on any of the disaster-causing similar typhoon retrieval model acquisition methods described above.

[0015] The present invention also provides a terminal, comprising: a processor and a memory; the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory to execute any of the above-described methods for obtaining disaster-causing similar typhoon retrieval models, or to execute the disaster-causing similar typhoon retrieval method as described above.

[0016] The present invention also provides a computer storage medium storing a computer program, wherein the computer program is executed by a processor using the method for obtaining a disaster-causing similar typhoon retrieval model as described above, or using the disaster-causing similar typhoon retrieval method as described above.

[0017] As described above, the present invention provides a method for obtaining a disaster-causing similar typhoon retrieval model, a retrieval method, a terminal, and a computer storage medium. This involves dividing target typhoons into first target typhoons and second target typhoons, and related typhoons into first related typhoons and second related typhoons. Based on the path similarity and meteorological field similarity between the first target typhoon and the first related typhoon, and combined with the fitting similarity between the second target typhoon and the second related typhoon, the weight parameters in the disaster-causing similar typhoon search model are optimized to obtain a weight-optimized model. Based on this weight-optimized model, disaster-causing similar typhoons corresponding to the typhoon to be searched are obtained. This improves the ability of disaster-causing similar typhoons to predict typhoons with abnormal paths, enhances the accuracy of disaster-causing similar typhoon searches, and improves the ability to assess the intensity of typhoon disasters. Attached Figure Description

[0018] Figure 1 The diagram shown is a flowchart of an embodiment of the method for obtaining disaster-causing similar typhoon retrieval models provided by the present invention.

[0019] Figure 2 The diagram shows a flowchart of step S800 during execution in an embodiment of the present invention.

[0020] Figure 3 The diagram shows a flowchart of step S804 during execution in an embodiment of the present invention.

[0021] Figure 4 The diagram shows the vector corresponding to the target typhoon OAB and the vector corresponding to related typhoons O'A'B' in this embodiment of the invention.

[0022] Vector distribution diagram;

[0023] Figure 5 The diagram shown is a flowchart of another embodiment of the method for obtaining disaster-causing similar typhoon retrieval models provided by the present invention;

[0024] Figure 6 The diagram shown is a flowchart of another embodiment of the method for obtaining disaster-causing similar typhoon retrieval models provided by the present invention;

[0025] Figure 7 The diagram shown is a flowchart of an embodiment of the disaster-causing similar typhoon retrieval method provided by the present invention;

[0026] Figure 8The results displayed are similar typhoons causing disasters obtained based on existing retrieval methods and those obtained using the method described in this invention.

[0027] The path distribution of typhoons that caused similar disasters;

[0028] Figure 9 The diagram shown is a structural schematic of a terminal provided in an embodiment of the present invention. Detailed Implementation

[0029] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, unless otherwise specified, the following embodiments and features described therein can be combined with each other.

[0030] To address the technical problems existing in the prior art, the present invention provides a method for obtaining a disaster-causing typhoon similarity retrieval model in the first aspect, for obtaining a disaster-causing typhoon similarity retrieval model adapted to the observation area.

[0031] Among them, the disaster-causing typhoons are those with similar disaster-causing capabilities and spatial distribution to the target typhoon; the typhoons are tropical cyclones, including six levels: tropical depression (TD), tropical storm (TS), severe tropical storm (STS), typhoon (TY), severe typhoon (STY), and super typhoon (Super TY).

[0032] Please see Figure 1 The diagram shows a flowchart illustrating the method for obtaining the disaster-causing similar typhoon retrieval model; as shown below. Figure 1 As shown, the method includes:

[0033] S200, determine the target typhoon in the observation area and the related typhoons corresponding to the target typhoon; obtain the spatial distribution information and meteorological field information of each path observation point in the target typhoon, and obtain the spatial distribution information and meteorological field information of each path observation point in the related typhoons.

[0034] The target typhoon is a historical typhoon to be retrieved within the observation area; the historical typhoon is a typhoon that has occurred in the observation area within a historical time period.

[0035] The related typhoons are those that have a spatiotemporal correlation with the target typhoon.

[0036] Optionally, typhoons that formed during the peak typhoon season in the observation area are selected as historical typhoons to exclude the influence of low-probability, non-dominant factors on the method. For example, for the Northwest Pacific observation area of ​​my country, typhoons that formed between June and November are selected as historical typhoons.

[0037] Optionally, the observation area ranges from 10°S to 60°N and from 90° to 180°E.

[0038] It should be noted that each target typhoon corresponds to multiple related typhoons, and each target typhoon and each related typhoon is a typhoon with a total observation duration of more than 36 hours.

[0039] The spatial distribution information of the path observation points refers to the location information of each observation point along the typhoon path.

[0040] The meteorological field information of the observation points along the typhoon path is the atmospheric circulation field information of each observation point at the corresponding observation time. The atmospheric circulation field information includes information on various meteorological elements at different altitude levels. Among them, the meteorological elements include one or more of humidity field, wind field, altitude field or other existing meteorological elements. For example, the atmospheric circulation field information includes a 500Ph geopotential height field or wind field, an 850hPa relative humidity field, etc.

[0041] Specifically, the system acquires a typhoon path dataset within the observation area; from the typhoon path dataset, it acquires the spatial distribution information of each path observation point in the target typhoon and the spatial distribution information of each path observation point in the related typhoons; and it acquires a typhoon meteorological field dataset within the observation area; from the typhoon meteorological field dataset, it acquires the meteorological field information corresponding to each path observation point in the target typhoon and the meteorological field information corresponding to each path observation point in the related typhoons.

[0042] Optionally, the method further includes, when performing step S200:

[0043] Based on the same encryption frequency, encrypted sampling is performed on the path observation points of the target typhoon and on the path observation points of the related typhoons; spatial distribution information and meteorological field information of the target typhoon at each path observation point are obtained, and spatial distribution information and meteorological field information of the related typhoons at each path observation point are obtained.

[0044] The time interval between two adjacent observation points on the target typhoon path is the same as the time interval between two adjacent observation points on the related typhoon path.

[0045] In one specific implementation, the target typhoon path and the related typhoon paths are both typhoon paths composed of 6-hour path observation points; using an interpolation method, the target typhoon path and the related typhoon paths are both densified into typhoon paths composed of 10-minute path observation points, so that the distribution of each path observation point can more accurately represent the path distribution of the typhoon.

[0046] The interpolation method includes linear interpolation, spline interpolation, or other existing interpolation methods.

[0047] S400, based on a preset path division method, each target typhoon is divided into a first target typhoon and a second target typhoon, and each related typhoon is divided into a first related typhoon and a second related typhoon; based on the correspondence between the target typhoons and the related typhoons, the first related typhoon corresponding to each first target typhoon is determined, and the second related typhoon corresponding to each second target typhoon is determined.

[0048] Specifically, for each path observation point in the target typhoon, the typhoon sub-paths corresponding to the n path observation points located at the end of the entire path are classified as the second target typhoon; the typhoon sub-paths corresponding to the other path observation points in the target typhoon are classified as the first target typhoon.

[0049] Similarly, for each path observation point in the relevant typhoon path, the typhoon sub-paths corresponding to the n path observation points located at the end of the entire path are classified as second relevant typhoons; the typhoon sub-paths corresponding to the other path observation points in the relevant typhoon are classified as first relevant typhoons.

[0050] Optionally, n can be any value from 3 to 5.

[0051] For a single target typhoon, the first related typhoon among its corresponding related typhoons is taken as the first related typhoon of the target typhoon, and the second related typhoon of each of its corresponding related typhoons is taken as the second related typhoon of the target typhoon.

[0052] S600, using the line similarity calculation method, obtain the path similarity between the first target typhoon and the corresponding first related typhoon, and using the field similarity calculation method, obtain the meteorological field similarity between the first target typhoon and the corresponding first related typhoon;

[0053] Wherein, the path similarity between the first target typhoon and the corresponding first related typhoon is the set of the shortest distance and distortion between each path observation point in the first target typhoon and the path observation point in the corresponding first related typhoon.

[0054] The meteorological field similarity is the similarity between the last path observation point of the first target typhoon and the last path observation point of the corresponding first related typhoon; the last path observation point is the last path observation point in the typhoon path.

[0055] In one specific embodiment, the method for obtaining the path similarity between the first target typhoon and the first related typhoon includes:

[0056] For the m-th path observation point in the first target typhoon i, calculate the distance between it and each path observation point in the first related typhoon j, and take the closest path observation point as the shortest path observation point of the m-th path observation point in the first related typhoon j. Remember the n-th path observation point; construct the vector MN pointing from the m-th path observation point to the n-th path observation point, and use it as the shortest path vector of the m-th path observation point in the first related typhoon j.

[0057] Perform the above steps for each path observation point in the first target typhoon i to obtain the nearest path observation point and the shortest path vector of each path observation point in the first related typhoon j.

[0058] The scalar values ​​of each of the shortest path vectors are averaged, and this average scalar value is taken as the shortest average distance between the first target typhoon i and the first related typhoon j, as follows:

[0059]

[0060] Furthermore, the average of each of the shortest path vectors is taken, and this average value is used as the average deviation between the first target typhoon i and the first related typhoon j, as follows:

[0061]

[0062] Based on the shortest path vectors and the average deviation, the average distortion between the first target typhoon i and the first related typhoon j is obtained.

[0063]

[0064] Based on the average distortion and shortest average distance between the first target typhoon i and the first related typhoon j, the path similarity between the first target typhoon i and the first related typhoon j is constructed as follows:

[0065]

[0066] In the formula, m represents the number of path observation points included in the path of the first target typhoon.

[0067] It should be noted that C ij The smaller the value, the higher the similarity between the two typhoon paths.

[0068] In one specific embodiment, the method for obtaining the meteorological field similarity between the first target typhoon path and the first related typhoon path includes:

[0069] Calculate the optimal similarity coefficient B between the meteorological field of the first target typhoon i and the meteorological field of the first historical typhoon j. ij The method is as follows:

[0070]

[0071]

[0072]

[0073] B ij =F ij ·V ij (8)

[0074] In the formula, x ik The variable value represents the value of each grid point within the range of the meteorological location. The average value of x is taken within the range of meteorological locations. This represents the average value of the variable at each grid point in the typhoon i meteorological field; Let m be the average value of the variable at each grid point in the meteorological field of typhoon j, and m be the number of grid points in the range of the meteorological field. ij The range of is (0,1], and when it is 1, the two samples completely overlap.

[0075] S800, based on the path similarity and meteorological field similarity between the first target typhoon and the corresponding first related typhoon, and combined with the fitting similarity between the second target typhoon and the corresponding second related typhoon, the weight parameters in the disaster-causing similar typhoon search model are optimized to obtain the final disaster-causing similar typhoon search model.

[0076] Specifically, when step S800 is executed, such as... Figure 2 As shown, it includes the following sub-steps:

[0077] S801, based on a preset weight interval, divides the value range of weight parameters in the disaster-causing typhoon similarity search model to obtain the initial model under different weight configurations;

[0078] The disaster-causing similar typhoon search model is a weighted model based on the path similarity and the meteorological field similarity.

[0079] In one specific embodiment, the disaster-causing similar typhoon search model is:

[0080]

[0081] In the formula, w represents the weight assigned to the similarity of the meteorological fields, and I i Search model for typhoons that cause disasters.

[0082] Based on a preset fixed window width interval, the path similarity C and the meteorological field similarity B are weighted and assigned to obtain a search model for disaster-causing similar typhoons under different weight configurations.

[0083] Optionally, the fixed window width interval is 0.1, that is, the weights of the path similarity C and the meteorological field similarity B are configured as shown in Table 1 below.

[0084] Table 1 shows the weight configurations for the path similarity C and the meteorological field similarity B (with a window width interval of 0.1).

[0085] Path similarity weight (1-w) Weather field similarity weight (w) 0.1 0.9 0.2 0.8 0.3 0.7 0.4 0.6 0.5 0.5 0.6 0.4 0.7 0.3 0.8 0.2 0.9 0.1

[0086] S802, based on the path similarity and meteorological field similarity between the first target typhoon and the corresponding first related typhoon, using each of the initial models, the total similarity between the first target typhoon and the corresponding first related typhoon under each weight configuration is obtained; the magnitudes of each total similarity are compared, and the optimal first related typhoon corresponding to each first target typhoon under each weight configuration is determined according to the comparison results;

[0087] The optimal first related typhoon is the first related typhoon with the highest total similarity to the first target typhoon under the current weight configuration.

[0088] Specifically, for the initial model under a single weight configuration, for example For any of the first target typhoons, perform the following operations respectively:

[0089] Input the path similarity and meteorological field similarity between the current first target typhoon and each of the corresponding first related typhoons into the current initial model; use the initial model to calculate the total similarity between the current first target typhoon and each of the corresponding first related typhoons;

[0090] Compare the magnitudes of each total similarity score and extract the maximum total similarity score.

[0091] The first related typhoon corresponding to the maximum total similarity is taken as the optimal first related typhoon corresponding to the current target typhoon, that is, the first related typhoon with the greatest disaster-causing similarity to the first target typhoon under the current weight configuration.

[0092] The above steps are performed on each of the first target typhoons to obtain the optimal first related typhoon corresponding to each of the first target typhoons under the current weight configuration.

[0093] For each of the weight configurations, step S802 is executed to obtain the optimal first related typhoon corresponding to each first target typhoon under each weight configuration.

[0094] S803, based on the correspondence between the first target typhoon and the optimal first related typhoon under each weight configuration, obtain the optimal second related typhoon corresponding to each second target typhoon;

[0095] Specifically, under a single weight configuration, for any of the first target typhoons, the following actions are performed:

[0096] The second target typhoon corresponding to the current first target typhoon is taken as the current second target typhoon;

[0097] Obtain the second relevant typhoon corresponding to the optimal first relevant typhoon;

[0098] Based on the correspondence between the first target typhoon and the optimal first related typhoon, the second related typhoon corresponding to the optimal first related typhoon is taken as the optimal second related typhoon of the current second target typhoon.

[0099] The above steps are performed on each of the first target typhoons to obtain the optimal second related typhoon corresponding to each of the second target typhoons under the current weight configuration.

[0100] For each of the aforementioned weight configurations, step S803 is executed to obtain the optimal second related typhoon corresponding to each second target typhoon under each weight configuration.

[0101] S804, using the path fitting similarity evaluation method, obtain the fitting similarity between the second target typhoon and the corresponding optimal second related typhoon under different weight configurations;

[0102] Specifically, under a single weight configuration, step S804, during execution, such as... Figure 3 As shown, it includes:

[0103] S804a, obtain the distance deviation between each second target typhoon and the corresponding optimal second related typhoon, and obtain the angle deviation between each second target typhoon and the corresponding optimal second related typhoon;

[0104] S804b, based on preset statistical features, perform comprehensive statistics on each distance deviation to obtain distance deviation feature values; and based on preset statistical features, perform comprehensive statistics on each angle deviation to obtain angle deviation feature values; and use the average of the distance deviation feature values ​​and the angle deviation feature values ​​as the fitting similarity between the second target typhoon and the corresponding optimal second related typhoon.

[0105] The above steps are performed for each weight configuration to obtain the fitting similarity between the second target typhoon and the corresponding optimal second related typhoon under each weight configuration.

[0106] In one specific embodiment, the distance deviation between the second target typhoon and the corresponding optimal second relevant typhoon is obtained as follows:

[0107] Based on the spatial distribution of the second target typhoon, a first vector is constructed, and based on the spatial distribution of the optimal second related typhoon, a second vector is constructed.

[0108] Subtract the second vector from the first vector, and use the vector difference as the distance deviation between the second target typhoon and the corresponding optimal second related typhoon.

[0109] For example, such as Figure 4 As shown, for target typhoon OAB, it includes first target typhoon OA and second target typhoon AB (24 hours after the tail of the target typhoon); for related typhoon O'A'B', it includes first related typhoon O'A' and second related typhoon A'B' (24 hours after the tail of the related typhoon); where the first related typhoon O'A' is the optimal first related typhoon of the first target typhoon OA, then the second related typhoon A'B' is determined to be the optimal second related typhoon of the second target typhoon AB; the vector corresponding to the second related typhoon A'B' is translated to point A to obtain a new vector A'C, then the vector BC is taken as the distance deviation between the second target typhoon and the corresponding optimal second related typhoon.

[0110] In one specific embodiment, the angular deviation between the second target typhoon and the corresponding optimal second relevant typhoon is obtained as follows:

[0111] Based on the spatial distribution of the second target typhoon, a first vector is constructed, and based on the spatial distribution of the optimal second related typhoon, a second vector is constructed; the angle between the first vector and the second vector is obtained, and this angle is used as the angular deviation between the second target typhoon and the corresponding optimal second related typhoon.

[0112] Still with Figure 4 For example, the angle between vector AB and vector AC is taken as the angular deviation between the second target typhoon and the corresponding optimal second related typhoon.

[0113] Optionally, the statistical features include the mean, median, and quartiles.

[0114] Specifically, under a single weight configuration, after obtaining the distance deviation between each second target typhoon and the corresponding optimal second related typhoon, the mean, median, and quartile of each distance deviation are extracted as feature values ​​of each distance deviation between the second target typhoon and the optimal second related typhoon; similarly, under a single weight configuration, after obtaining the angular deviation between each second target typhoon and the corresponding optimal second related typhoon, the mean, median, and quartile of each angular deviation are extracted as feature values ​​of each angular deviation between the second target typhoon and the optimal second related typhoon.

[0115] The quartiles include the lower quartile q1 (P25) and the upper quartile q3 (P75).

[0116] In a more specific embodiment, under different weight configurations, the distance deviation characteristic value and angle deviation characteristic value between the second target typhoon and the optimal second related typhoon are shown in Table 2 below.

[0117] Table 2 shows the characteristic values ​​of distance deviation and angle deviation between each second target typhoon and the optimal second related typhoon.

[0118]

[0119] S805, compare the fitting similarity values ​​of each, and take the weight configuration corresponding to the maximum fitting similarity value as the optimal weight configuration, thereby obtaining the final disaster-causing similar typhoon search model.

[0120] In another embodiment, the method for obtaining the disaster-causing similar typhoon retrieval model, before executing step S800, is as follows: Figure 5 As shown, it also includes:

[0121] S700, perform standardization processing on the path similarity and meteorological field similarity between each of the first target typhoons and the corresponding first related typhoons, and perform subsequent steps based on the standardized path similarity and meteorological field similarity.

[0122] Specifically, in order to eliminate the dimensional differences between different parameters, the path similarity C and the meteorological field similarity B are standardized respectively.

[0123] In one specific implementation, the path similarity between the first target typhoon i and the corresponding first related typhoon j is standardized as follows:

[0124]

[0125] In the formula, For the standardized path similarity, C ij C represents the path similarity before standardization.imax C represents the maximum path similarity between each target typhoon and its corresponding related typhoons. imin This represents the minimum path similarity between each target typhoon and its corresponding related typhoon.

[0126] Furthermore, the meteorological field similarity between the first target typhoon i and the corresponding first related typhoon j is standardized as follows:

[0127]

[0128] In the formula, For the similarity of the standardized meteorological fields, B ij B represents the path similarity before standardization. imax B represents the maximum meteorological field similarity between each target typhoon and its corresponding related typhoons; imin This represents the minimum meteorological field similarity between each target typhoon and its corresponding related typhoon.

[0129] In another embodiment, the method for obtaining the disaster-causing similar typhoon retrieval model, before executing step S200, is as follows: Figure 6 As shown, it also includes:

[0130] S100, among the historical typhoons in the observation area, select the target typhoon and select the related typhoons corresponding to the target typhoon;

[0131] Specifically, the historical typhoons in the observation area within a certain time period are obtained, and several typhoons are randomly selected from these historical typhoons as the target typhoons.

[0132] Based on the path distribution of the target typhoon and the path distribution of other historical typhoons, the spatial relationship between the target typhoon and other historical typhoons is extracted, and it is detected whether the spatial relationship meets the preset spatial conditions. If so, other historical typhoons that meet the spatial relationship are regarded as related typhoons of the target typhoon.

[0133] In one specific embodiment, for a single target typhoon, the method for extracting the spatial positional relationship between the target typhoon and other historical typhoons includes: selecting any typhoon from other historical typhoons, extracting the distance between each path observation point on the path of the typhoon and any path observation point in the current target typhoon, detecting whether each distance is less than a preset distance threshold, and if so, then taking the typhoon as a related typhoon of the current target typhoon.

[0134] Optionally, the distance threshold is 5 degrees.

[0135] In another embodiment, where the relevant typhoon is a typhoon that is spatially and temporally related to the target typhoon, the method for selecting the relevant typhoon further includes:

[0136] Based on the observation time of each path observation point in the target typhoon and the observation time of each path observation point in other historical typhoons, the time difference between the target typhoon and the other historical typhoons in terms of observation time is extracted; it is detected whether the time difference is less than a preset time difference threshold. If so, the other historical typhoons are regarded as related typhoons of the target typhoon.

[0137] Specifically, for a single target typhoon, the method for extracting the time difference between the target typhoon and other historical typhoons in terms of observation time includes: selecting any typhoon from the other historical typhoons, extracting the time difference between the observation time of each path observation point on the path of the typhoon and the observation time of any path observation point in the current target typhoon; detecting whether each time difference is less than a preset time difference threshold, if so, then taking the typhoon as a related typhoon of the current target typhoon; performing this step for each target typhoon to obtain the related typhoons corresponding to each target typhoon.

[0138] Optionally, the time difference threshold is 60 days.

[0139] To address the technical problems existing in the prior art, the present invention provides a method for retrieving disaster-causing typhoons with similarity in disaster-causing characteristics from historical typhoons within an observation area.

[0140] Please see Figure 7 The diagram below illustrates the flowchart of the method for retrieving disaster-causing typhoons similar to those described above; Figure 7 As shown, the method includes the following steps:

[0141] S10, select a target typhoon within the observation area, and select a related typhoon corresponding to the target typhoon;

[0142] In this embodiment, the target typhoon is a historical typhoon or a currently occurring typhoon within the observation area to be retrieved.

[0143] S20: Obtain spatial distribution information and meteorological field information of the target typhoon at each path observation point, and obtain spatial distribution information and meteorological field information of each related typhoon at each path observation point;

[0144] S30, using the line similarity calculation method, obtain the path similarity between the target typhoon and each related typhoon, and using the field similarity calculation method, obtain the meteorological field similarity between the target typhoon and each related typhoon;

[0145] In this embodiment, the method for obtaining the path similarity between the target typhoon and each related typhoon is the same as in the above embodiments, and will not be repeated here; similarly, the method for obtaining the meteorological field similarity between the target typhoon and each related typhoon is the same as in the above embodiments, and will not be repeated here.

[0146] S40, based on the path similarity and meteorological field similarity between the target typhoon and each related typhoon, the total similarity between the target typhoon and each related typhoon is obtained using a disaster-causing similar typhoon search model, and the similar typhoon corresponding to the maximum total similarity is taken as the disaster-causing similar typhoon of the target typhoon.

[0147] The disaster-causing similar typhoon search model is a model obtained by the disaster-causing similar typhoon retrieval model obtained in the above embodiments.

[0148] Specifically, the path similarity and meteorological field similarity between the target typhoon and each related typhoon are input into the disaster-causing similar typhoon search model to obtain the total similarity between the target typhoon and each related typhoon.

[0149] Compare the magnitudes of each total similarity and extract the maximum value among them; use the relevant typhoon corresponding to the maximum value as the disaster-causing similar typhoon of the target typhoon.

[0150] To verify the beneficial effects of the method described in this invention, taking Typhoon Jangpi (No. 201525) in the western Pacific region as an example, the corresponding disaster-causing similar typhoons were obtained using both existing similar typhoon retrieval methods (without considering meteorological field factors) and the disaster-causing similar typhoon retrieval method provided in this invention. Please refer to [link / reference]. Figure 8The diagram shows the path distribution of similar typhoons causing disasters obtained using existing retrieval methods and those obtained using the method described in this invention; wherein, the similar typhoon causing disasters obtained using existing retrieval methods is No. 199725 (blue), and the similar typhoon causing disasters obtained using the method described in this invention is No. 200623 (red). As shown in the figure, although Typhoon No. 199725 (blue) and the target typhoon Jangpi had a high degree of similarity in their paths and were closer in spatial location, while Typhoon No. 200623 (red) was farther away from Jangpi, after reaching the path observation point at 00:00 on October 18, 2015 (hereinafter referred to as the "target observation point"), Typhoon Jangpi's path underwent a significant turn, moving from northwest to north-northeast. Typhoon No. 199725, on the other hand, continued to move northwest, meaning its predicted path (blue dashed line) deviated significantly from Jangpi's path, with a deviation of 331 km between the predicted path and Jangpi's (24 hours after the target observation point). Typhoon No. 200623, however, had a similar path turn to Jangpi, with a deviation of only 23.6 km between its predicted path (red dashed line) and Jangpi's (24 hours after the target observation point). Based on this, it can be seen that the disaster-causing similar typhoon retrieval method accurately predicted the path turning point of Typhoon Jangpi.

[0151] Furthermore, to verify the effectiveness of the disaster-causing similar typhoons obtained by the method described in this invention in predicting the intensity of disaster losses, taking Typhoon Saomai (No. 200610) and Typhoon Lekima (No. 201912) as examples, historical typhoons with disaster records from 1985 to 2019 were retrieved. The disaster-causing similar typhoons corresponding to Saomai, namely Typhoon No. 201514, and the disaster-causing similar typhoons corresponding to Lekima, namely Typhoon No. 198709, were obtained using the method described in this invention. The intensity of disaster losses of each target typhoon and its corresponding disaster-causing similar typhoons was compared, and the comparison results are shown in Table 3 below.

[0152] Table 3. Comparison of damage caused by Typhoons Saoma and Lekima with corresponding typhoons of similar severity.

[0153]

[0154] As shown in the table, for the four provinces with the most severe disaster losses caused by Typhoon Soudelor (target typhoon 1) and its similar typhoon Soudelor, the correlation coefficients for the affected population, relocated population, affected area, and direct economic losses are as high as 0.93. If we look at the provinces that were jointly affected, the correlation coefficients for the above four loss indicators for Zhejiang, Fujian, and Jiangxi are also as high as 0.91.

[0155] Regarding Typhoon Lekima (target typhoon 2) and its similar typhoon No. 198709, the two typhoons are quite similar in terms of the area of ​​damage they caused.

[0156] As can be seen from the above, the disaster-causing similar typhoons obtained based on the method of the present invention can better indicate the turning changes in the typhoon path compared with the disaster-causing similar typhoons obtained by existing methods. Moreover, the obtained disaster-causing similar typhoons also have better indicative significance in terms of disaster intensity and magnitude forecasting and judgment. Therefore, the disaster warning accuracy of the method of the present invention is higher.

[0157] To address the problems existing in the prior art, the present invention also provides a terminal in a second aspect, please refer to [link / reference]. Figure 9 A schematic diagram of the structure of the terminal described in this invention is shown; as follows: Figure 9 As shown, the terminal 5 includes a memory 51 and a processor 52 connected to each other; the memory 51 is used to store computer programs, and the processor 52 is used to execute the computer programs stored in the memory, so that when the terminal is executed, it can implement the steps in the disaster-causing similar typhoon retrieval model acquisition method or the disaster-causing similar typhoon retrieval method described in the above embodiment.

[0158] Optionally, the number of memories can be one or more, and the number of processors can be one or more.

[0159] Optionally, the processor in the terminal loads one or more instructions corresponding to application processes into the memory according to the steps in the disaster-sustaining typhoon retrieval model acquisition method or the disaster-sustaining typhoon retrieval method described in the above embodiments, and the processor runs the application stored in the memory, thereby realizing the functions in the disaster-sustaining typhoon retrieval model acquisition method or the disaster-sustaining typhoon retrieval method described above, which will not be elaborated here.

[0160] It should be noted that memory includes, but is not limited to, random access memory (RAM), and may also include non-volatile memory, such as at least one disk storage device. Similarly, processors can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0161] This invention also provides a computer-readable storage medium storing a computer program thereon, which, when called by a processor, implements the disaster-causing similar typhoon retrieval model acquisition method or the disaster-causing similar typhoon retrieval method described above.

[0162] A computer-readable storage medium can be a tangible device capable of holding and storing instructions used by an instruction execution device. Computer-readable storage media can be, for example, (but not limited to) electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, and mechanical encoding devices.

[0163] The computer-readable program described herein can be downloaded from a computer-readable storage medium to various computing / processing devices, or downloaded via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network, to an external computer or external storage device. A network adapter card or network interface in each computing / processing device receives computer-readable program instructions from the network and forwards these instructions to the computer-readable storage medium in the respective computing / processing device.

[0164] In summary, the disaster-causing similar typhoon retrieval model acquisition method, retrieval method, terminal, and computer storage medium provided by this invention obtain the path similarity and meteorological field similarity between the target typhoon and related typhoons. Based on this path similarity and meteorological field similarity, and combined with the fitting similarity between the subsequent paths of the target typhoon and related typhoons, the weight parameters in the disaster-causing similar typhoon search model are optimized to obtain a weight-optimized model. Based on the weight-optimized disaster-causing similar typhoon search model, disaster-causing similar typhoons corresponding to the typhoon to be searched are obtained. This improves the ability of disaster-causing similar typhoons to predict typhoons with abnormal paths, improves the accuracy of disaster-causing similar typhoon search, and improves the ability to judge the intensity of typhoon disasters. Therefore, it has an important indicative role in the judgment and analysis of typhoon disasters. In addition, by using a traversal method to search the weight parameters of the model, the amount of data processing during the weight parameter search process can be reduced, improving both model performance and model acquisition efficiency.

[0165] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A method for obtaining a disaster-causing typhoon similarity retrieval model, characterized in that, include: The spatial distribution information and meteorological field information of each observation point along the path of the target typhoon are obtained, as well as the spatial distribution information and meteorological field information of each observation point along the path of related typhoons; wherein, the related typhoons are historical typhoons in the observation area that are associated with the target typhoon. Based on a preset path division method, the target typhoon is divided into a first target typhoon and a second target typhoon, and each related typhoon is divided into a first related typhoon and a second related typhoon. The division method of the first and second target typhoons among the target typhoons includes: for each path observation point in the target typhoon, the typhoon sub-paths corresponding to the n path observation points located at the end of the entire path are divided into second target typhoons; the typhoon sub-paths corresponding to the other path observation points in the target typhoon are divided into first target typhoons; based on the target typhoon and the... The correspondence between related typhoons is determined by identifying the first related typhoons corresponding to the first target typhoon and the second related typhoons corresponding to the second target typhoon. The method for dividing the first and second target typhoons among the target typhoons includes: for each path observation point in the path of the related typhoons, the typhoon sub-paths corresponding to the n path observation points located at the end of the entire path are classified as second related typhoons; the typhoon sub-paths corresponding to the other path observation points in the related typhoons are classified as first related typhoons; where n is any value from 3 to 5. The path similarity between the first target typhoon and the corresponding first related typhoon is obtained by using the line similarity calculation method, and the meteorological field similarity between the first target typhoon and the corresponding first related typhoon is obtained by using the field similarity calculation method. Based on the path similarity and meteorological field similarity between the first target typhoon and its corresponding first related typhoon, and combined with the fitting similarity between the second target typhoon and its corresponding second related typhoon, the weight parameters in the disaster-causing similar typhoon search model are optimized to obtain the final disaster-causing similar typhoon search model. The optimization configuration method for the weight parameters in the disaster-causing similar typhoon search model includes: dividing the value range of the weight parameters in the disaster-causing similar typhoon search model based on a preset weight interval to obtain initial models under different weight configurations; based on the path similarity and meteorological field similarity between the first target typhoon and its corresponding first related typhoon, using each of the initial models, obtaining the first weight parameter under each weight configuration. The overall similarity between a target typhoon and its corresponding first related typhoon is calculated. The magnitudes of these overall similarities are compared, and the optimal first related typhoon for each target typhoon under each weight configuration is determined based on the comparison results. Based on the correspondence between the first target typhoon and the optimal first related typhoon under each weight configuration, the optimal second related typhoon for each second target typhoon is determined. A path fitting similarity evaluation method is used to obtain the fitting similarity between the second target typhoon and its corresponding optimal second related typhoon under different weight configurations. The magnitudes of these fitting similarities are compared, and the weight configuration corresponding to the maximum fitting similarity is taken as the optimal weight configuration, thereby obtaining the final disaster-causing similar typhoon search model. The disaster-causing similar typhoon search model is a weighted model based on the path similarity and the meteorological field similarity.

2. The method for obtaining a disaster-causing similar typhoon retrieval model according to claim 1, characterized in that, Before optimizing the weight parameters in the disaster-causing similar typhoon search model, the following steps are also included: The path similarity and meteorological field similarity between the first target typhoon and the corresponding first related typhoon are standardized, and subsequent steps are performed based on the standardized path similarity and meteorological field similarity.

3. The method for obtaining a disaster-causing similar typhoon retrieval model according to claim 1, characterized in that, Also includes: Among the historical typhoons within the observation area, the target typhoon and the related typhoons corresponding to the target typhoon are selected.

4. The method for obtaining a disaster-causing similar typhoon retrieval model according to claim 3, characterized in that, The methods for obtaining the relevant typhoons corresponding to the target typhoon include: Based on the path distribution of the target typhoon and the path distribution of other historical typhoons, the spatial relationship between the target typhoon and other historical typhoons is extracted, and it is detected whether the spatial relationship meets the preset spatial conditions. If so, other historical typhoons that meet the spatial relationship are considered as related typhoons of the target typhoon.

5. The method for obtaining a disaster-causing similar typhoon retrieval model according to claim 1, characterized in that, For a single weight configuration, the method for obtaining the fitting similarity between the second target typhoon and the corresponding optimal second related typhoon includes: Obtain the distance deviation between each second target typhoon and the corresponding optimal second related typhoon, and obtain the angle deviation between each second target typhoon and the corresponding optimal second related typhoon; Based on preset statistical characteristics, the distance deviations are integrated to obtain distance deviation feature values; and based on preset statistical characteristics, the angle deviations are integrated to obtain angle deviation feature values. The mean of the distance deviation feature value and the angle deviation feature value is used as the fitting similarity between the second target typhoon and the corresponding optimal second related typhoon.

6. The method for obtaining a disaster-causing similar typhoon retrieval model according to claim 5, characterized in that, The method for obtaining the distance deviation feature value includes: After obtaining the distance deviation between each second target typhoon and the corresponding optimal second related typhoon, the mean, median and quartile of each distance deviation are extracted as feature values ​​of each distance deviation between the second target typhoon and the optimal second related typhoon. And, the method for obtaining the angular deviation feature value includes: After obtaining the angular deviation between each second target typhoon and the corresponding optimal second related typhoon, the mean, median and quartile of each angular deviation are extracted as the characteristic values ​​of each angular deviation between the second target typhoon and the optimal second related typhoon.

7. A method for retrieving disaster-causing typhoons with similar characteristics, characterized in that, include: Select a target typhoon within the observation area, and select a related typhoon corresponding to the target typhoon; Acquire spatial distribution information and meteorological field information of the target typhoon at each path observation point, and acquire spatial distribution information and meteorological field information of each related typhoon at each path observation point; The path similarity between the target typhoon and each related typhoon is obtained by using the line similarity calculation method, and the meteorological field similarity between the target typhoon and each related typhoon is obtained by using the field similarity calculation method. as well as, Based on the path similarity and meteorological field similarity between the target typhoon and each related typhoon, the total similarity between the target typhoon and each related typhoon is obtained by using the disaster-causing similar typhoon search model. The similar typhoon corresponding to the maximum total similarity is taken as the disaster-causing similar typhoon of the target typhoon. The disaster-causing similar typhoon search model is a model obtained based on the disaster-causing similar typhoon retrieval model acquisition method as described in any one of claims 1 to 6.

8. A terminal, characterized in that, include: Processor and memory; The memory is used to store computer programs, and the processor is used to execute the computer programs stored in the memory to perform the disaster-causing similar typhoon retrieval model acquisition method as described in any one of claims 1 to 6, or to perform the disaster-causing similar typhoon retrieval method as described in claim 7.

9. A computer storage medium storing a computer program, characterized in that, The computer program is executed by a processor using the method for obtaining disaster-causing similar typhoon retrieval models as described in any one of claims 1 to 6, or using the method for retrieving disaster-causing similar typhoons as described in claim 7.

Images