Dynamic typhoon similarity matching method based on meteorological structure similarity

Abstract

The invention discloses a dynamic typhoon similarity matching method based on meteorological structure similarity, and the method comprises the steps: obtaining the feature data of a current typhoon and a historical typhoon, and the feature data comprise typhoon path data and typhoon meteorological data; the method comprises the following steps: selecting three continuous real-time points of a current typhoon and three continuous typhoon point data in historical typhoons, respectively calculating the similarity of gas field element fields on the three real-time points and the typhoon points, and sequentially screening a plurality of historical typhoons which are most similar to the current typhoon; according to the screened historical typhoon paths most similar to the current typhoon, the minimum value of the accumulated distance between the historical typhoon and the current typhoon path is calculated, the historical typhoon corresponding to the minimum value of the accumulated distance is the historical typhoon most similar to the current typhoon, and the historical typhoon is successfully matched with the current typhoon. According to the method, the similarity of typhoon paths is not simply considered, comprehensive consideration is carried out in combination with the typhoon meteorological field, historical typhoons with multiple similar elements are matched, and formulating of typhoon coping strategies can be better helped.

Description

Technical Field

[0001] The invention relates to the field of meteorological research, and in particular to a dynamic typhoon similarity matching method based on meteorological structure similarity. Background Art

[0002] Most of the existing typhoon similarity matching methods are based on distance measurement and simple dynamic time warping to perform similarity matching of typhoon paths. However, these methods do not consider complex meteorological factors and are heavily dependent on historical typhoon path data. The similar typhoons matched by these methods have poor accuracy, which seriously affects the accuracy of typhoon meteorological research. Summary of the invention

[0003] In view of the above-mentioned deficiencies in the prior art, the present invention provides a dynamic typhoon similarity matching method based on meteorological structure similarity, a new method for typhoon similarity matching that combines the meteorological characteristics of pressure field and wind field and performs historical path similarity to improve the accuracy of typhoon prediction.

[0004] In order to achieve the above-mentioned object of the invention, the technical solution adopted by the present invention is:

[0005] A dynamic typhoon similarity matching method based on meteorological structure similarity is provided, which comprises the following steps:

[0006] S1: Obtain characteristic data of current typhoon and historical typhoon, the characteristic data includes typhoon path data and typhoon meteorological data;

[0007] S2: Select three consecutive real-time points of the current typhoon and three consecutive typhoon points in the historical typhoons, calculate the similarity of the aura elements between the three real-time points and the typhoon points, and select several historical typhoons that are most similar to the current typhoon in turn;

[0008] S3: Based on the paths of several historical typhoons that are most similar to the current typhoon, the minimum cumulative distance between the historical typhoon and the current typhoon path is calculated. The historical typhoon corresponding to the minimum cumulative distance is the historical typhoon that is most similar to the current typhoon, and the historical typhoon is successfully matched with the current typhoon.

[0009] Further, step S2 includes:

[0010] S21: according to the characteristic data of the current typhoon and the historical typhoons, the three consecutive real-time point data (C1, C2, C3) of the current typhoon are selected, and the three consecutive typhoon point data (H1, H2, H3) in the historical typhoons are selected, and the time interval between the three typhoon point data (H1, H2, H3) is the same as the time interval between the three consecutive real-time point data (C1, C2, C3);

[0011] S22: Calculate the similarity of the meteorological element field between the first actual point C1 of the current typhoon and the first typhoon point H1 in the historical typhoons, and select m1 historical typhoon points H1 that are most similar to the actual point C1 according to the similarity of the meteorological element field, and the m1 historical typhoon points H1 correspond to the m1 most similar historical typhoons;

[0012] S23: Calculate the similarity of meteorological element fields between the second typhoon point H2 in the most similar m1 historical typhoons and the second actual point C2 of the current typhoon, and select the m2 historical typhoon points H2 that are most similar to the actual point C2 according to the similarity of meteorological element fields. The most similar m2 historical typhoon points H2 correspond to the most similar m2 historical typhoons, and m2 <m1;

[0013] S24: Calculate the similarity of meteorological element fields between the third typhoon point H3 of the most similar m2 historical typhoons and the third actual point C3 of the current typhoon, and select the m3 historical typhoon points H3 that are most similar to the actual point C3 according to the similarity of meteorological element fields. The most similar m3 historical typhoon points H3 correspond to the most similar m3 historical typhoons, and m3 <m2。

[0014] Furthermore, the calculation method of the meteorological element field similarity between the typhoon point and the actual point includes:

[0015] A21: Get the longitude and latitude corresponding to the typhoon point and the actual point (lon a ,lat a )、(lon b ,lat b ), a is the actual point, b is the typhoon point, (lon a ,lat a ) is the longitude and latitude of the actual point, (lon b ,lat b ) are the longitude and latitude of the typhoon point;

[0016] A22: Calculate the longitude and latitude translation Δlon and Δlat between the typhoon point and the actual point;

[0017] Δlon=lon a -lon b ,Δlat=lat a -lat b ;

[0018] A23: The meteorological element field M of typhoon point b b By moving the translation amounts Δlon and Δlat to the position aligned with the actual point a, the meteorological element field M after alignment b The longitude and latitude corresponding to the grid point (i, j) in is lon b ′(i,j), latb ′(i,j):

[0019] lon b ′(i,j)=lon b (i,j)+Δlon,lat b ′(i,j)=lat b (i,j)+Δlat;

[0020] Among them, lon b (i,j), lat b (i, j) are the initial longitude and latitude of the grid point (i, j), i is the discrete position index of the grid point in the latitude direction, and j is the discrete position index of the grid point in the longitude direction;

[0021] A24: Meteorological element field M b Each grid point in the grid is aligned with the meteorological element field M of the current typhoon. a Within the same longitude and latitude range;

[0022] A25: Calculation of meteorological element field M based on the Laplace operator of two-dimensional pressure field b and meteorological element field M a The pressure field weight of each grid point in the , considering the contribution of different regions in the pressure field, the similarity of the pressure field between the typhoon point and the actual point is calculated by the weighted structural similarity index method;

[0023] A26: According to the meteorological elements field M b and meteorological element field M a The wind field weight of each grid point is calculated based on the distance from each grid point to the typhoon center point, and the horizontal wind speed similarity and vertical wind speed similarity are calculated based on the wind field weight to obtain the wind field similarity.

[0024] A27: Use the pressure field similarity and wind field similarity to calculate the similarity of meteorological element fields between typhoon points and actual points.

[0025] Further, step A25 includes:

[0026] A251: Calculate the Laplace operator of the two-dimensional pressure field based on the two-dimensional pressure field P(i,j) of the grid point in the meteorological element field

[0027]

[0028] Where x is the east-west direction of the air pressure P, It represents the rate of change of air pressure P in the east-west direction, represents the rate of change of air pressure P along the north-south direction, y is the north-south direction of air pressure P, Δx and Δy are the longitude and latitude distances between the grid points of the two-dimensional air pressure field, respectively;

[0029] A252: Laplace operator based on two-dimensional pressure field Calculate the pressure field weight w for each grid point lap (i,j);

[0030]

[0031] A253: Using the pressure field weight w at each grid point lap (i, j), and considering the contribution of different regions in the pressure field, the pressure field similarity SSIM (P C ,P H );

[0032]

[0033] Among them, μ C , μ H are the current typhoon pressure field P C (i, j) and the pressure field P of historical typhoons H The brightness similarity of (i,j) at the grid point (i,j), σ C , σ H are the current typhoon pressure field P C (i, j) and the pressure field P of historical typhoons H The structural similarity of (i,j) at the grid point (i,j), σ CH is the pressure field P C (i, j) and the pressure field P H where is the covariance between (i, j), N is the number of grid points in the meteorological element field, and c1 and c2 are correction constants for the similarity of the pressure field.

[0034] Further, step A26 includes:

[0035] A261: Calculate the wind field weight ω(i,j) of each grid point according to the distance r(i,j) from each grid point in the meteorological element field to the typhoon center point;

[0036]

[0037] A262: Calculate the wind field similarity between the typhoon point and the actual point according to the wind field weight ω(i,j) of each grid point. The wind field similarity includes horizontal wind speed similarity and vertical wind speed similarity.

[0038] Horizontal wind speed similarity SSIM u(v C ,v H ) is calculated as:

[0039]

[0040] in, are the horizontal wind speed of the current typhoon and horizontal wind speeds of historical typhoons The brightness similarity at the grid point (i,j), are the horizontal wind speed of the current typhoon and horizontal wind speeds of historical typhoons The similarity of the structure at the grid point (i,j), Horizontal wind speed With horizontal wind speed The covariance between them, c′1 and c′2 are both correction constants for the similarity of horizontal wind speed;

[0041] Similarly, calculate the vertical wind speed similarity SSIM v (u C ,u H );

[0042]

[0043] in, are the vertical wind speed of the current typhoon and vertical wind speed of historical typhoons The brightness similarity at the grid point (i,j), are the vertical wind speed of the current typhoon and vertical wind speed of historical typhoons The similarity of the structure at the grid point (i,j), is the vertical wind speed Vertical wind speed The covariance between them, c″1, c″2 are correction constants for the similarity of vertical wind speed;

[0044] A263: Based on vertical wind speed similarity SSIM v (u C ,u H ) and horizontal wind speed similarity SSIM u (v C ,v H ) Calculate wind field similarity SSIM V ;

[0045] SSIM V =SSIM v (u C ,u H )+SSIMu (v C ,v H ).

[0046] Furthermore, the meteorological element field similarity SSIM between the actual point and the typhoon point total The calculation method is:

[0047]

[0048] Among them, α and β are the weight coefficients of pressure field similarity and wind field similarity, respectively.

[0049] Further, step S3 includes:

[0050] S31: according to the selected most similar m3 historical typhoons, the cumulative distance between the paths of the m3 historical typhoons and the path of the current typhoon is calculated according to the dynamic time warping algorithm, and the minimum cumulative distance between the m3 historical typhoons and the path of the current typhoon is selected;

[0051] The calculation method for the minimum cumulative distance between historical typhoons and the current typhoon path is:

[0052] Define the current typhoon path as C = [C1, C2, ..., C n ], the path of historical typhoons is H = [H1, H2, ..., H m ], n is the number of path points of the current typhoon, m is the number of path points of historical typhoons, C n is the nth path point of the current typhoon, C m is the mth path point of the historical typhoon;

[0053] Calculate the minimum cumulative distance DTW(C,H) between the historical typhoon path and the current typhoon path;

[0054]

[0055] Among them, e is the number of the current typhoon path point, f is the number of the historical typhoon path point, C e is the e-th path point of the current typhoon, C f is the f-th path point of the historical typhoon, π is the alignment path between the path points of the current typhoon and the path points of the historical typhoon, (e,f)∈π indicates that the e-th path point of the current typhoon is aligned with the f-th path point of the historical typhoon, d(C e ,H f ) is the distance between the alignment paths;

[0056] S32: The historical typhoon corresponding to the minimum cumulative distance is the historical typhoon most similar to the current typhoon, and the historical typhoon is successfully matched with the current typhoon.

[0057] The beneficial effects of the present invention are as follows: the present invention aims to solve the problem that the existing typhoon similarity technology only considers path similarity but does not sufficiently analyze the similarity of pressure fields and wind fields. By combining the global similarity calculation of pressure fields and wind fields and using the dynamic time warping (DTW) algorithm to perform global matching of typhoon paths, the method can gradually screen out historical typhoons that are most similar to the current typhoon, and make a more accurate prediction of the path and intensity of the current typhoon.

[0058] The present invention uses a weighted structural similarity index algorithm and a dynamic time warping algorithm to compare the similarity between the current typhoon and the historical typhoon. First, the current typhoon and historical typhoon data are obtained, and then the current typhoon meteorological field and the historical typhoon meteorological field data are aligned according to the longitude and latitude coordinates. The Laplace operator and Gaussian decrease are used to construct the weight function of the typhoon pressure field and wind field. According to the weight function, multiple historical typhoons that are most similar to the current typhoon meteorological field are gradually found, and then the dynamic warping algorithm is used to find the most similar historical typhoon. Compared with the existing typhoon similarity comparison method, this method does not simply consider the similarity of the typhoon path but comprehensively considers the typhoon meteorological field, matches the historical typhoons with similar multiple elements, and can better help formulate typhoon response strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 This is a flow chart of the dynamic typhoon similarity matching method based on meteorological structure similarity. DETAILED DESCRIPTION

[0060] The specific implementation modes of the present invention are described below so that those skilled in the art can understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation modes. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations utilizing the concept of the present invention are protected.

[0061] like Figure 1 As shown, a dynamic typhoon similarity matching method based on meteorological structure similarity includes the following steps:

[0062] S1: Obtain the characteristic data of current typhoons and historical typhoons. The characteristic data include typhoon path data and typhoon meteorological data. Typhoon path data can be obtained through the typhoon network of the Central Meteorological Observatory, and typhoon meteorological data can be obtained from the Global Data Assimilation System (GDAS);

[0063] S2: Select three consecutive real-time points of the current typhoon and three consecutive typhoon points in the historical typhoons, calculate the similarity of the aura elements between the three real-time points and the typhoon points, and select several historical typhoons that are most similar to the current typhoon in turn;

[0064] Step S2 specifically includes:

[0065] According to the characteristic data of the current typhoon and historical typhoons, the three most recent continuous real-time point data of the current typhoon (C1, C2, C3) are selected, and the three continuous typhoon point data of the historical typhoon (H1, H2, H3) are selected, and the time interval between the three typhoon point data (H1, H2, H3) is the same as the time interval between the three continuous real-time point data (C1, C2, C3);

[0066] Calculate the similarity of the meteorological element field between the first actual point C1 of the current typhoon and the first typhoon point H1 in the historical typhoons, and select the 30 historical typhoon points H1 that are most similar to the actual point C1 based on the similarity of the meteorological element field. The 30 historical typhoon points H1 correspond to the 30 most similar historical typhoons. The higher the similarity of the meteorological element field, the higher the similarity between the actual point and the typhoon point. The 30 most similar historical typhoon points are the top 30 historical typhoon points in terms of meteorological element field similarity.

[0067] Calculate the similarity of meteorological element fields between the second typhoon point H2 of the 30 most similar historical typhoons and the second actual point C2 of the current typhoon, and select the 20 historical typhoon points H2 that are most similar to the actual point C2 based on the similarity of meteorological element fields. The 20 most similar historical typhoon points H2 correspond to the 20 most similar historical typhoons.

[0068] Calculate the similarity of meteorological element fields between the third typhoon point H3 of the 20 most similar historical typhoons and the third actual point C3 of the current typhoon, and select the 10 historical typhoon points H3 that are most similar to the actual point C3 based on the similarity of meteorological element fields. The 10 most similar historical typhoon points H3 correspond to the 10 most similar historical typhoons.

[0069] The similarity calculation method of meteorological element fields between typhoon points and actual points is:

[0070] Get the longitude and latitude (lon) corresponding to the typhoon point and the actual point a ,lat a )、(lon b ,lat b ), a is the actual point, b is the typhoon point, (lon a ,lat a ) is the longitude and latitude of the actual point, (lon b ,lat b ) are the longitude and latitude of the typhoon point;

[0071] Calculate the longitude and latitude translation Δlon and Δlat between the typhoon point and the actual point;

[0072] Δlon=lon a-lon b ,Δlat=lat a -lat b ;

[0073] The meteorological element field M of typhoon point b b By moving the translation amounts Δlon and Δlat to the position aligned with the actual point a, the meteorological element field M after alignment b The longitude and latitude corresponding to the grid point (i, j) in is lon b ′(i,j), lat b ′(i,j):

[0074] lon b ′(i,j)=lon b (i,j)+Δlon,lat b ′(i,j)=lat b (i,j)+Δlat;

[0075] Among them, lon b (i,j), lat b (i, j) are the initial longitude and latitude of the grid point (i, j), i is the discrete position index of the grid point in the latitude direction, and j is the discrete position index of the grid point in the longitude direction;

[0076] The meteorological element field M b Each grid point in the grid is aligned with the meteorological element field M of the current typhoon. a Within the same longitude and latitude range;

[0077] When analyzing the pressure field, ridges and troughs are very important structural features. They correspond to high-pressure and low-pressure areas of air pressure, and have an important impact on the evolution of weather systems, especially typhoon paths. The present invention locates these significant features by calculating the second-order derivative of the pressure field. The wind field near the center of the typhoon has the greatest impact on the intensity and path of the typhoon, while the wind field in the edge area plays an important role in the interaction between the path adjustment and the external wind field. Therefore, the change of the overall wind field, especially near the center of the typhoon, is a key factor affecting the behavior of the typhoon.

[0078] In order to accurately identify the ridges and grooves under the air pressure field, the present invention introduces the second-order derivative of the air pressure field (i.e., the Laplace operator). The second-order derivative represents the acceleration of the air pressure change, which can help us determine the location area of ​​the ridges and grooves.

[0079] According to the two-dimensional pressure field P(i,j) of the grid point in the meteorological element field, calculate the Laplace operator of the two-dimensional pressure field

[0080]

[0081] Where x is the east-west direction of the air pressure P, It represents the rate of change of air pressure P in the east-west direction, It represents the rate of change of air pressure P along the north-south direction, y is the north-south direction of air pressure P, Δx and Δy are the longitude and latitude distances between the grid points of the two-dimensional pressure field respectively; the larger the Laplace operator value, the more significant the pressure change, and the greater the impact on the evolution of the typhoon path.

[0082] The pressure field is based on the weight distribution function of the Laplace operator. The value of the Laplace operator reflects the acceleration of the change of the pressure field. The larger its absolute value, the greater the possibility that the point is located on the ridge line and trough line. Since the ridge line and trough line have a great influence on the typhoon path, they should be given a higher weight.

[0083] According to the Laplace operator of the two-dimensional pressure field Calculate the pressure field weight w for each grid point lap (i,j);

[0084]

[0085] Based on the identification of ridges and troughs, the similarity of the pressure field is calculated using an improved weighted structural similarity index algorithm, which considers the contribution of different regions in the pressure field in a weighted manner.

[0086] Using the pressure field weight w at each grid point lap (i, j), and considering the contribution of different regions in the pressure field, the pressure field similarity SSIM (P C ,P H );

[0087]

[0088] Among them, μ C , μ H are the current typhoon pressure field P C (i, j) and the pressure field P of historical typhoons H The brightness similarity of (i,j) at the grid point (i,j), σ C , σ H are the current typhoon pressure field P C (i, j) and the pressure field P of historical typhoons H The structural similarity of (i,j) at the grid point (i,j), σ CH is the pressure field P C (i, j) and the pressure field P HThe covariance between (i, j), N is the number of grid points in the meteorological element field, c1 and c2 are correction constants for the similarity of the pressure field to prevent the denominator from being zero or numerically unstable;

[0089] The above formula takes into account the similarity of brightness (mean), contrast (variance) and structure (covariance) of each grid point in the pressure field in a weighted manner, and combines the weights of important areas such as ridges and grooves to measure the global similarity of the two pressure fields. The larger the numerator and the smaller the denominator, the higher the similarity of the two pressure fields.

[0090] The weight function of the wind field is calculated, and the Gaussian decreasing weight assignment strategy from the center to the edge is adopted. This method emphasizes the importance of the wind field in the central area of ​​the typhoon, reduces the influence of the edge area, and can flexibly control the decreasing rate of the weight.

[0091] According to the distance r(i,j) from each grid point in the meteorological element field to the typhoon center, calculate the wind field weight ω(i,j) of each grid point;

[0092]

[0093] The wind field similarity between the typhoon point and the actual point is calculated according to the wind field weight ω(i,j) of each grid point. The wind field similarity includes horizontal wind speed similarity and vertical wind speed similarity.

[0094] Horizontal wind speed similarity SSIM u (v C ,v H ) is calculated as:

[0095]

[0096] in, are the horizontal wind speed of the current typhoon and horizontal wind speeds of historical typhoons The brightness similarity at the grid point (i,j), are the horizontal wind speed of the current typhoon and horizontal wind speeds of historical typhoons The similarity of the structure at the grid point (i,j), Horizontal wind speed With horizontal wind speed The covariance between them, c′1 and c′2 are both correction constants for the similarity of horizontal wind speed;

[0097] Similarly, calculate the vertical wind speed similarity SSIM v (u C ,u H );

[0098]

[0099] in, are the vertical wind speed of the current typhoon and vertical wind speed of historical typhoons The brightness similarity at the grid point (i,j), are the vertical wind speed of the current typhoon and vertical wind speed of historical typhoons The similarity of the structure at the grid point (i,j), is the vertical wind speed Vertical wind speed The covariance between them, c″1, c″2 are correction constants for the similarity of vertical wind speed;

[0100] Based on the vertical wind speed similarity SSIM v (u C ,u H ) and horizontal wind speed similarity SSIM u (v C ,v H ) Calculate wind field similarity SSIM V ;

[0101] SSIM V =SSIM v (u C ,u H )+SSIM u (v C ,v H );

[0102] Calculate the meteorological element field similarity SSIM between the actual point and the typhoon point total ;

[0103]

[0104] Among them, α and β are the weight coefficients of pressure field similarity and wind field similarity, respectively.

[0105] S3: Based on the paths of several historical typhoons that are most similar to the current typhoon, the minimum cumulative distance between the historical typhoon and the current typhoon path is calculated. The historical typhoon corresponding to the minimum cumulative distance is the historical typhoon that is most similar to the current typhoon, and the historical typhoon is successfully matched with the current typhoon. Step S3 specifically includes:

[0106] Based on the 10 most similar historical typhoons selected, the cumulative distance between the paths of the 10 historical typhoons and the path of the current typhoon is calculated using the dynamic time warping algorithm, and the minimum cumulative distance between the 10 historical typhoons and the path of the current typhoon is selected;

[0107] The calculation method for the minimum cumulative distance between historical typhoons and the current typhoon path is:

[0108] Define the current typhoon path as C = [C1, C2, ..., C n ], the path of historical typhoons is H = [H1, H2, ..., H m ], n is the number of path points of the current typhoon, m is the number of path points of historical typhoons, C n is the nth path point of the current typhoon, C m is the mth path point of the historical typhoon;

[0109] Calculate the minimum cumulative distance DTW(C,H) between the historical typhoon path and the current typhoon path;

[0110]

[0111] Among them, e is the number of the current typhoon path point, f is the number of the historical typhoon path point, C e is the e-th path point of the current typhoon, C f is the f-th path point of the historical typhoon, π is the alignment path between the path points of the current typhoon and the path points of the historical typhoon, (e,f)∈π indicates that the e-th path point of the current typhoon is aligned with the f-th path point of the historical typhoon, d(C e ,H f ) is the distance between the alignment paths;

[0112] The historical typhoon corresponding to the minimum cumulative distance is the most similar historical typhoon to the current typhoon, and the historical typhoon is successfully matched with the current typhoon. When dealing with the similarity comparison of typhoon paths, dynamic time warping can well deal with the misalignment problem on the speed, length or time axis, and effectively find the best alignment method for comparison.

[0113] The present invention aims to solve the problem that the existing typhoon similarity technology only considers path similarity but does not sufficiently analyze the similarity of pressure field and wind field. By combining the global similarity calculation of pressure field and wind field and using the dynamic time warping (DTW) algorithm to perform global matching of typhoon paths, this method can gradually screen out historical typhoons that are most similar to the current typhoon, and make a more accurate prediction of the path and intensity of the current typhoon.

[0114] The present invention uses a weighted structural similarity index algorithm and a dynamic time warping algorithm to compare the similarity between the current typhoon and the historical typhoon. First, the current typhoon and historical typhoon data are obtained, and then the current typhoon meteorological field and the historical typhoon meteorological field data are aligned according to the longitude and latitude coordinates. The Laplace operator and Gaussian decrease are used to construct the weight function of the typhoon pressure field and wind field. According to the weight function, multiple historical typhoons that are most similar to the current typhoon meteorological field are gradually found, and then the dynamic warping algorithm is used to find the most similar historical typhoon. Compared with the existing typhoon similarity comparison method, this method does not simply consider the similarity of the typhoon path but comprehensively considers the typhoon meteorological field, matches the historical typhoons with similar multiple elements, and can better help formulate typhoon response strategies.

Claims

1. A dynamic typhoon similarity matching method based on meteorological structure similarity, characterized in that: The following steps are involved: S1: Obtain characteristic data of current typhoon and historical typhoon, the characteristic data includes typhoon path data and typhoon meteorological data; S2: Select three consecutive real-time points of the current typhoon and three consecutive typhoon points in the historical typhoons, calculate the similarity of the aura elements between the three real-time points and the typhoon points, and select several historical typhoons that are most similar to the current typhoon in turn; S3: Based on the paths of several historical typhoons that are most similar to the current typhoon, the minimum cumulative distance between the historical typhoon and the current typhoon path is calculated. The historical typhoon corresponding to the minimum cumulative distance is the historical typhoon that is most similar to the current typhoon, and the historical typhoon is successfully matched with the current typhoon.

2. The dynamic typhoon similarity matching method based on meteorological structure similarity according to claim 1 is characterized in that: The step S2 comprises: S21: according to the characteristic data of the current typhoon and the historical typhoons, the three consecutive real-time point data (C1, C2, C3) of the current typhoon are selected, and the three consecutive typhoon point data (H1, H2, H3) in the historical typhoons are selected, and the time interval between the three typhoon point data (H1, H2, H3) is the same as the time interval between the three consecutive real-time point data (C1, C2, C3); S22: Calculate the similarity of the meteorological element field between the first actual point C1 of the current typhoon and the first typhoon point H1 in the historical typhoons, and select m1 historical typhoon points H1 that are most similar to the actual point C1 according to the similarity of the meteorological element field, and the m1 historical typhoon points H1 correspond to the m1 most similar historical typhoons; S23: Calculate the similarity of meteorological element fields between the second typhoon point H2 in the most similar m1 historical typhoons and the second actual point C2 of the current typhoon, and select the m2 historical typhoon points H2 that are most similar to the actual point C2 according to the similarity of meteorological element fields. The most similar m2 historical typhoon points H2 correspond to the most similar m2 historical typhoons, and m2 <m1; S24: Calculate the similarity of meteorological element fields between the third typhoon point H3 of the most similar m2 historical typhoons and the third actual point C3 of the current typhoon, and select the m3 historical typhoon points H3 that are most similar to the actual point C3 according to the similarity of meteorological element fields. The most similar m3 historical typhoon points H3 correspond to the most similar m3 historical typhoons, and m3 <m2。 3. The dynamic typhoon similarity matching method based on meteorological structure similarity according to claim 2 is characterized in that: The method for calculating the similarity of the meteorological element field between the typhoon point and the actual point includes: A21: Get the longitude and latitude corresponding to the typhoon point and the actual point (lon a ,lat a )、(lon b ,lat b ), a is the actual point, b is the typhoon point, (lon a ,lat a ) is the longitude and latitude of the actual point, (lon b ,lat b ) are the longitude and latitude of the typhoon point; A22: Calculate the longitude and latitude translation Δlon and Δlat between the typhoon point and the actual point; Δlon=lon a -lon b ,Δlat=lat a -years b ; A23: The meteorological element field M of typhoon point b b By moving the translation amounts Δlon and Δlat to the position aligned with the actual point a, the meteorological element field M after alignment b The longitude and latitude corresponding to the grid point (i, j) in is lon b ′(i,j), lat b ′(i,j): lon b ′(i,j)=lon b (i,j)+Δlon,lat b ′(i,j)=lat b (i,j)+Δlat; Among them, lon b (i,j), lat b (i, j) are the initial longitude and latitude of the grid point (i, j), i is the discrete position index of the grid point in the latitude direction, and j is the discrete position index of the grid point in the longitude direction; A24: Meteorological element field M b Each grid point in the grid is aligned with the meteorological element field M of the current typhoon. a Within the same longitude and latitude range; A25: Calculation of meteorological element field M based on the Laplace operator of two-dimensional pressure field b and meteorological element field M a The pressure field weight of each grid point in the , considering the contribution of different regions in the pressure field, the similarity of the pressure field between the typhoon point and the actual point is calculated by the weighted structural similarity index method; A26: According to the meteorological elements field M b and meteorological element field M a The wind field weight of each grid point is calculated based on the distance from each grid point to the typhoon center point, and the horizontal wind speed similarity and vertical wind speed similarity are calculated based on the wind field weight to obtain the wind field similarity. A27: Use the pressure field similarity and wind field similarity to calculate the similarity of meteorological element fields between typhoon points and actual points.

4. The dynamic typhoon similarity matching method based on meteorological structure similarity according to claim 3 is characterized in that: The step A25 comprises: A251: Calculate the Laplace operator of the two-dimensional pressure field based on the two-dimensional pressure field P(i,j) of the grid point in the meteorological element field Where x is the east-west direction of the air pressure P, It represents the rate of change of air pressure P in the east-west direction, represents the rate of change of air pressure P along the north-south direction, y is the north-south direction of air pressure P, Δx and Δy are the longitude and latitude distances between the grid points of the two-dimensional air pressure field, respectively; A252: Laplace operator based on two-dimensional pressure field Calculate the pressure field weight w for each grid point lap (i,j); A253: Using the pressure field weight w at each grid point lap (i, j), and considering the contribution of different regions in the pressure field, the pressure field similarity SSIM (P C ,P H ); Among them, μ C , μ H are the current typhoon pressure field P C (i, j) and the pressure field P of historical typhoons H The brightness similarity of (i,j) at the grid point (i,j), σ C , σ H are the current typhoon pressure field P C (i, j) and the pressure field P of historical typhoons H The structural similarity of (i,j) at the grid point (i,j), σ CH is the pressure field P C (i, j) and the pressure field P H where is the covariance between (i, j), N is the number of grid points in the meteorological element field, and c1 and c2 are correction constants for the similarity of the pressure field.

5. The dynamic typhoon similarity matching method based on meteorological structure similarity according to claim 4 is characterized in that: The step A26 comprises: A261: Calculate the wind field weight ω(i,j) of each grid point according to the distance r(i,j) from each grid point in the meteorological element field to the typhoon center point; A262: Calculate the wind field similarity between the typhoon point and the actual point according to the wind field weight ω(i,j) of each grid point. The wind field similarity includes horizontal wind speed similarity and vertical wind speed similarity. Horizontal wind speed similarity SSIM u (v C ,v H ) is calculated as: in, are the horizontal wind speeds of the current typhoon and horizontal wind speeds of historical typhoons The brightness similarity at the grid point (i,j), are the horizontal wind speed of the current typhoon and horizontal wind speeds of historical typhoons The similarity of the structure at the grid point (i,j), Horizontal wind speed With horizontal wind speed The covariance between them, c′1 and c′2 are both correction constants for the similarity of horizontal wind speed; Similarly, calculate the vertical wind speed similarity SSIM v (u C ,u H ); in, are the vertical wind speed of the current typhoon and vertical wind speed of historical typhoons The brightness similarity at the grid point (i,j), are the vertical wind speed of the current typhoon and vertical wind speed of historical typhoons The similarity of the structure at the grid point (i,j), is the vertical wind speed Vertical wind speed The covariance between them, c″1, c″2 are correction constants for the similarity of vertical wind speed; A263: Based on vertical wind speed similarity SSIM v (u C ,u H ) and horizontal wind speed similarity SSIM u (v C ,v H ) Calculate wind field similarity SSIM V ; SSIM V =SSIM v (u C ,u H )+SSIM u (v C ,v H )。 6. The dynamic typhoon similarity matching method based on meteorological structure similarity according to claim 5, characterized in that: The meteorological element field similarity SSIM between the actual point and the typhoon point total The calculation method is: Among them, α and β are the weight coefficients of pressure field similarity and wind field similarity, respectively.

7. The dynamic typhoon similarity matching method based on meteorological structure similarity according to claim 2 is characterized in that: The step S3 comprises: S31: according to the selected most similar m3 historical typhoons, the cumulative distance between the paths of the m3 historical typhoons and the path of the current typhoon is calculated according to the dynamic time warping algorithm, and the minimum cumulative distance between the m3 historical typhoons and the path of the current typhoon is selected; The calculation method for the minimum cumulative distance between historical typhoons and the current typhoon path is: Define the current typhoon path as C = [C1, C2, ..., C n ], the path of historical typhoons is H = [H1, H2, ..., H m ], n is the number of path points of the current typhoon, m is the number of path points of historical typhoons, C n is the nth path point of the current typhoon, C m is the mth path point of the historical typhoon; Calculate the minimum cumulative distance DTW(C,H) between the historical typhoon path and the current typhoon path; Among them, e is the number of the current typhoon path point, f is the number of the historical typhoon path point, C e is the e-th path point of the current typhoon, C f is the f-th path point of the historical typhoon, π is the alignment path between the path points of the current typhoon and the path points of the historical typhoon, (e,f)∈π indicates that the e-th path point of the current typhoon is aligned with the f-th path point of the historical typhoon, d(C e ,H f ) is the distance between the alignment paths; S32: The historical typhoon corresponding to the minimum cumulative distance is the historical typhoon most similar to the current typhoon, and the historical typhoon is successfully matched with the current typhoon.

Images