Tropical cyclone path screening method, system, equipment, medium and product

Abstract

The invention discloses a tropical cyclone path screening method, system and device, a medium and a product, and relates to the field of ocean and weather forecast, and the method comprises the steps: selecting a tropical cyclone sample, and determining a first path similar typhoon list of the tropical cyclone sample based on a tropical cyclone path similarity method; based on a comparative learning tropical cyclone similarity evaluation model, determining a second path similar typhoon list of the tropical cyclone sample; the comparative learning tropical cyclone similarity evaluation model is obtained by training a training data set; the training data set comprises a CMA tropical cyclone optimal path data set and ERA5 reanalysis data information; according to the first path similar typhoon list and the second path similar typhoon list, a tropical cyclone list similar to the tropical cyclone sample path is determined, and an optimal tropical cyclone path is screened from the tropical cyclone list. And the misjudgment condition is avoided.

Description

Technical Field

[0001] The present application relates to the fields of ocean and meteorological forecasting, and particularly to a method, system, device, medium and product for screening tropical cyclone paths. Background Art

[0002] Research on the progress of tropical cyclone similarity, that is, how tropical cyclones evolve and exhibit similar behavioral patterns, has always been a major focus of meteorology. Similarity is manifested in multiple aspects such as the trajectories, intensities, and structures of tropical cyclones. Usually, several key aspects are considered when studying tropical cyclone similarity.

[0003] 1. Climatic characteristics.

[0004] Geographical distribution: Tropical cyclones follow certain paths, usually affected by sea surface temperature, the Coriolis effect, and the prevailing wind patterns in the region.

[0005] Seasonal trends: Most tropical cyclones occur in specific seasons in the Pacific and Atlantic basins (such as from June to November in the western Pacific). The storm formation, development, and trajectories during this period usually exhibit similar patterns. Formation conditions: The formation of tropical cyclones requires specific conditions, such as warm sea water, high humidity, and low vertical wind shear. Similar conditions result in storms with similar characteristics.

[0006] 2. Trajectory similarity and intensity similarity.

[0007] Analyze the historical paths of typhoons to determine the trajectories where typhoons repeatedly occur or the areas where storms are likely to form and intensify. As is well known, the waters east of the Philippines are important birthplaces of tropical cyclones. Intensity similarity, especially manifested as the rapid intensification of tropical cyclones near the coast, usually results in a 30 hPa (hectopascal) drop in the central pressure within 24 hours, causing a rapid increase in tropical intensity. Usually, such tropical cyclones will cause significant catastrophic impacts along the coast.

[0008] 3. Impact similarity.

[0009] By identifying similar patterns among past typhoons, better warning systems can be developed to predict storm paths, intensities, and affected areas. Tropical cyclones have similar developments or intensities in terms of rainfall, wind disasters, and storm surges, and typhoons with similar characteristics will have similar impacts.

[0010] The above three aspects are from a macroscopic perspective of tropical cyclone similarity. Specifically in terms of the paths of tropical cyclones, their trajectories are affected by multiple factors of the atmosphere and the ocean. The following are the recognized main factors.

[0011] 1. Sea surface temperature.

[0012] The energy of a tropical cyclone comes from warm seawater. Higher sea surface temperatures (generally above 26°C) provide the necessary heat and moisture for tropical cyclones, making them stronger and more persistent. The presence of warm water also affects the movement and duration of cyclones. Although the sea surface temperature does not directly affect the direction of the cyclone, it indirectly affects the intensity and development of the cyclone, thus influencing the movement track of the tropical cyclone.

[0013] 2. Wind shear.

[0014] Wind shear refers to the change in wind speed and direction at different heights. Strong wind shear (a strong difference in wind speed and direction between the lower and upper levels of the atmosphere) can weaken or destroy the structure of a tropical cyclone. Low wind shear is favorable for the formation and strengthening of cyclones. On the contrary, high wind shear will weaken the cyclone, possibly changing its path or causing it to dissipate. The direction and intensity of wind shear affect the movement of the cyclone.

[0015] 3. Steering winds (prevailing winds).

[0016] Steering winds are the winds in the middle and upper levels of the atmosphere (usually 500 mb - 700 mb), which guide the movement of tropical cyclones along their paths. These winds change over time, geographical location, and seasons. The most important steering winds for tropical cyclones are the easterly winds in the tropics, which usually push the cyclones towards the west. However, changes in steering winds, such as those caused by the subtropical jet or large high-pressure systems, can cause the cyclone to turn northward or eastward.

[0017] 4. Pressure systems.

[0018] Tropical cyclones usually move along the edges of high- and low-pressure systems. The position and intensity of these pressure systems have a significant impact on the track of the cyclone. High-pressure systems are barriers that force the cyclone to move around them. They can direct the cyclone in a specific direction or completely block its advance. Low-pressure systems attract tropical cyclones, causing them to change direction, accelerate, or change their paths. The interaction between high- and low-pressure systems is crucial in determining the movement of the cyclone.

[0019] 5. Terrain.

[0020] When a tropical cyclone encounters land, the interaction between the storm and the land surface will significantly change its path and intensity. Due to the lack of warm seawater, friction with the land surface, and the reduction in moisture supply, cyclones usually weaken when they make landfall. If the cyclone approaches large mountain ranges (such as the Himalayas, the Andes), deflection, weakening, or even dissipation may occur. Mountains can block the movement of the cyclone or cause the cyclone to turn along both sides of the mountains. Coastal landforms such as bays or peninsulas can change the path of the cyclone by concentrating wind patterns and storm surges, usually enhancing the impact of the cyclone in certain areas.

[0021] 6. Mid-latitude troughs and jets.

[0022] Once a tropical cyclone starts moving away from the tropics, the positions of mid-latitude troughs and jets can affect the movement of the tropical cyclone. When a tropical cyclone enters the mid-latitudes, it interacts with the troughs and jet streams, which usually causes the cyclone to turn towards higher latitudes.

[0023] 7. Atmospheric instability and convection.

[0024] Cyclones are driven by convection, which is the upward movement of warm, moist air. Atmospheric instability promotes the development of convection, providing power for the cyclone. The interaction between the cyclone and unstable atmospheric conditions affects the movement and development of the cyclone. Areas where air rises cause changes in wind patterns, thus altering the cyclone's track.

[0025] 8. Structure and size of the cyclone.

[0026] The structure of the cyclone, including its size (i.e., the radius of the storm) and wind field distribution, affects the cyclone's response to steering winds and pressure systems. Larger cyclones are more affected by large-scale steering winds and may take longer to respond to changes in environmental conditions, while smaller cyclones may be more unstable and may change direction more quickly.

[0027] Two tropical cyclones with similar paths can, to a certain extent, reflect the similarities in the environmental conditions and evolution processes of tropical cyclones, and reflect the equivalence of the combined effects of many influencing factors. It is widely used in operations such as tropical cyclone forecasting and typhoon storm surge forecasting. Taking typhoon storm surge forecasting as an example, finding historical tropical cyclones with paths similar to the sample tropical cyclone, consulting the development of historical tropical cyclones and the actual situation of coastal water levels and storm surges caused by them has very important reference value for predicting the storm surge that the current sample tropical cyclone may trigger.

[0028] Through the analysis of existing patents and relevant literature, the existing similarity of tropical cyclones can be mainly reflected in single indicators such as path similarity and intensity similarity. In path similarity, basically, it starts from graphical similarity, and by giving quantitative indicators, the degree of graphical similarity is measured. However, the problem with the above method is that it ignores that a tropical cyclone, as a weather-scale system, its trajectory development and change are affected by a variety of oceanic and meteorological factors. Its graph (trajectory) is only the final manifestation, and these complex influencing factors are the root causes. Tropical cyclones with similar paths may have very different possible intensities, moving speeds, and their circulation systems. If the tropical cyclone path is determined only based on the degree of path similarity, misjudgment may occur. Therefore, multiple factors need to be comprehensively considered in tropical cyclone similarity. Summary of the Invention

[0029] The objective of this application is to provide a tropical cyclone path screening method, system, device, medium and product, so as to solve the problem that the tropical cyclone paths determined only by the path similarity are prone to misjudgment.

[0030] To achieve the above objective, this application provides the following solutions.

[0031] In the first aspect, this application provides a tropical cyclone path screening method, including the following steps.

[0032] Select a tropical cyclone sample, and based on the tropical cyclone path similarity method, determine the first path-similar typhoon list of the tropical cyclone sample.

[0033] Based on the contrastive learning tropical cyclone similarity evaluation model, determine the second path-similar typhoon list of the tropical cyclone sample; the contrastive learning tropical cyclone similarity evaluation model is obtained by training with a training data set; the training data set includes the CMA tropical cyclone best track data set and the ERA5 reanalysis data; the CMA tropical cyclone best track data set includes the time, intensity, latitude, longitude and central minimum pressure of the tropical cyclone; the ERA5 reanalysis data includes surface elements and meteorological elements at different pressure levels; the surface elements include sea level pressure and sea surface wind; the meteorological elements include geopotential height, wind, air temperature, vorticity and divergence.

[0034] According to the first path-similar typhoon list and the second path-similar typhoon list, determine the tropical cyclone list that is path-similar to the tropical cyclone sample, and screen the best tropical cyclone path from the tropical cyclone list.

[0035] In the second aspect, this application provides a tropical cyclone path screening system, including the following modules.

[0036] The first path-similar typhoon list determination module is used to select a tropical cyclone sample and determine the first path-similar typhoon list of the tropical cyclone sample based on the tropical cyclone path similarity method.

[0037] The second path-similar typhoon list determination module is used to determine the second path-similar typhoon list of the tropical cyclone sample based on the contrastive learning tropical cyclone similarity evaluation model; the contrastive learning tropical cyclone similarity evaluation model is obtained by training with a training data set; the training data set includes the CMA tropical cyclone best track data set and the ERA5 reanalysis data; the CMA tropical cyclone best track data set includes the time, intensity, latitude, longitude and central minimum pressure of the tropical cyclone; the ERA5 reanalysis data includes surface elements and meteorological elements at different pressure levels; the surface elements include sea level pressure and sea surface wind; the meteorological elements include geopotential height, wind, air temperature, vorticity and divergence.

[0038] The optimal tropical cyclone path screening module is used to determine a list of tropical cyclones with paths similar to the tropical cyclone sample path according to the first list of typhoons with similar paths and the second list of typhoons with similar paths, and screen the optimal tropical cyclone path from the list of tropical cyclones.

[0039] In a third aspect, the present application provides a computer device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the computer program to implement the above-mentioned tropical cyclone path screening method.

[0040] In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the above-mentioned tropical cyclone path screening method is implemented.

[0041] In a fifth aspect, the present application provides a computer program product, including a computer program, and when the computer program is executed by a processor, the above-mentioned tropical cyclone path screening method is implemented.

[0042] According to the specific embodiments provided by the present application, the present application has the following technical effects: The present application first determines the first list of typhoons with similar paths based on the tropical cyclone path similarity method. Secondly, based on the contrastive learning tropical cyclone similarity evaluation model trained by the training data set constructed from the CMA tropical cyclone best track data set and the ERA5 reanalysis data, the second list of typhoons with similar paths of the tropical cyclone sample is determined. The training data set includes different influencing factors of tropical cyclones to characterize the interaction relationship between the ocean and the atmosphere, so that the list of tropical cyclones with paths similar to the tropical cyclone sample path determined according to the first list of typhoons with similar paths and the second list of typhoons with similar paths is more accurate, and thus the optimal tropical cyclone path can be selected. The present application not only considers the trajectory graph, but also considers the interaction relationship between the ocean and the atmosphere, comprehensively considers various factors to evaluate the similarity degree of tropical cyclones, and avoids misjudgment. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0044] Figure 1 It is a schematic flowchart of a tropical cyclone path screening method provided by an embodiment of the present application.

[0045] Figure 2 Schematic diagram of a low-pressure high-wind area provided by an embodiment of the present application.

[0046] Figure 3 Schematic diagram of the correspondence between tropical cyclone numbers - time and result E provided by an embodiment of the present application.

[0047] Figure 4 Schematic diagram of the structure of a computer device provided by an embodiment of the present application. Detailed implementation manners

[0048] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0049] To make the objectives, features, and advantages of the present application more obvious and understandable, the present application will be further described in detail below in conjunction with the accompanying drawings and specific implementation manners.

[0050] The embodiments of the present application provide a method for screening tropical cyclone paths, including the following steps.

[0051] S1: Select a tropical cyclone sample, and based on the tropical cyclone path similarity method, determine the first list of typhoons with similar paths to the tropical cyclone sample.

[0052] S2: Based on the contrastive learning tropical cyclone similarity evaluation model, determine the second list of typhoons with similar paths to the tropical cyclone sample; the contrastive learning tropical cyclone similarity evaluation model is obtained by training with a training data set; the training data set includes the CMA tropical cyclone best track data set and the ERA5 reanalysis data; the CMA tropical cyclone best track data set includes the time, intensity, latitude, longitude, and central minimum pressure of the tropical cyclone; the ERA5 reanalysis data includes surface elements and meteorological elements at different pressure levels; the surface elements include sea-level pressure and sea surface wind; the meteorological elements include geopotential height, wind, air temperature, vorticity, and divergence.

[0053] S3: According to the first list of typhoons with similar paths and the second list of typhoons with similar paths, determine the list of tropical cyclones with paths similar to the tropical cyclone sample, and screen the best tropical cyclone path from the list of tropical cyclones.

[0054] In practical applications, the present application selects a tropical cyclone as a sample; based on the calculation results of traditional path similarity (such as through buffer analysis, etc.), that is, the tropical cyclone path similarity method, determines the first list of typhoons with similar paths for the selected sample tropical cyclone; based on the calculation results of the contrastive learning tropical cyclone similarity evaluation model, determines the second list of typhoons with similar paths for the selected sample tropical cyclone; and takes the intersection of the first list of typhoons with similar paths and the second list of typhoons with similar paths as the final tropical cyclone list.

[0055] Among them, the tropical cyclone path similarity method can be the tropical cyclone path similarity method based on the Fréchet distance, the tropical cyclone buffer zone spatial analysis method, or the method described in the method and storage medium for determining the degree of tropical cyclone path similarity.

[0056] The present application also designs a method for selecting, organizing, and processing from the training data set to generate a training data set, and based on the contrastive learning model, conducts model training and verification. After the training results pass the verification and evaluation, the model is applied to actual operations, introducing the synoptic features of the subtropical high and low-pressure systems.

[0057] In an exemplary embodiment, before S2, it further includes: processing the training data set to determine the processed training data set.

[0058] In an exemplary embodiment, the background of data set processing is closely related to the application scenario of tropical meteorological similarity. The analysis of tropical cyclone similarity focuses on the similarity for a period of time before and after the landfall of a tropical cyclone (such as 72 hours before landfall), rather than the entire path starting from the generation source area, which can be understood as local similarity. Based on this, processing the training data set to determine the processed training data set specifically includes the following steps.

[0059] S11: Based on the tropical cyclone serial number, process each tropical cyclone in the CMA tropical cyclone best track data set one by one to form a path point data set; the processing includes deleting records with the sub-center identifier.

[0060] S12: Based on the tropical cyclone serial number, extract the 500hPa geopotential height grid data, sea-level pressure data, 10m wind U-component data, and 10m wind V-component data from the ERA5 reanalysis data according to time; where U is the horizontal wind speed component; V is the vertical wind speed component.

[0061] S13: Determine the subtropical high isosurface according to the 500hPa geopotential height grid data.

[0062] S14: Determine the tropical cyclone low-pressure area based on the sea-level air pressure data, the 10-meter wind U-component data, and the 10-meter wind V-component data.

[0063] S15: Integrate the path point data set, the subtropical high isosurface, and the tropical cyclone low-pressure area to construct a processed training data set.

[0064] In practical applications, the China Meteorological Administration (CMA) tropical cyclone best track data set comes from the data of the CMA Tropical Cyclone Data Center. The CMA tropical cyclone best track data set includes the position and intensity of tropical cyclones in the waters of the Northwest Pacific (north of the equator, west of 180° east longitude) every 6 hours since 1949, including time, intensity, latitude, longitude, and the lowest central air pressure, etc. Among them, the time is Coordinated Universal Time (UTC).

[0065] The ERA5 reanalysis data comes from the European Centre for Medium-Range Weather Forecasts and includes ERA5 hourly data on pressure levels from 1940 to present and ERA5 hourly data on single levels from 1940 to present. Specifically, it includes meteorological elements such as geopotential height, wind, air temperature, vorticity, and divergence at different pressure levels since 1940. The surface elements include sea-level air pressure and sea surface wind (10 meters), etc.

[0066] In an exemplary embodiment, S13 may include the following steps.

[0067] S21: Using the contour line with a value greater than 5880 gpm as a condition, perform contour line tracing on the 500 hPa geopotential height grid data, and use the Marching Squares algorithm to determine the first contour line result.

[0068] S22: Conduct a preliminary screening on the first contour line result to determine the contour line result after preliminary screening.

[0069] S23: Use the Douglas–Peucker algorithm to simplify the contour line result after preliminary screening to determine the contour line result after simplification.

[0070] S24: Connect each point in the contour line result after simplification in sequence to generate the subtropical high surface data; the subtropical high surface data is the subtropical high isosurface.

[0071] In an exemplary embodiment, S14 may include the following steps.

[0072] S31: Perform contour tracing on the sea level pressure data in the range of 910 hPa to 1000 hPa at intervals of 5 hPa, and use the Marching Squares algorithm to determine the second contour result.

[0073] S32: Determine the low-pressure area of the 990 hPa isobar based on the second contour result.

[0074] S33: Generate gridded wind speed data based on the 10-meter wind U-component data and the 10-meter wind V-component data.

[0075] S34: Based on the gridded wind speed data, extract the contour area with wind force above level 6; the contour area with wind force above level 6 is the strong wind area with wind force above level 6.

[0076] S35: Determine the low-pressure strong wind area based on the intersection area of the low-pressure area and the strong wind area; the low-pressure strong wind area is the tropical cyclone low-pressure area.

[0077] In an exemplary embodiment, S32 may include the following steps.

[0078] S41: Process each item of the second contour result according to the threshold condition to determine the processed contour result.

[0079] S42: Use the Douglas–Peucker algorithm to simplify the processed contour result to determine the simplified contour result, and select the low-pressure area of the 990 hPa isobar based on the simplified contour result.

[0080] Taking the northwest Pacific region as an example, further illustrate the generation and processing process of the training dataset.

[0081] First, it is necessary to determine that: 1. Tropical cyclones are affected by the subtropical high, and the spatial distribution of the subtropical high has an obvious impact on the movement of tropical cyclones.

[0082] 2. Low-pressure systems can attract tropical cyclones, causing them to change direction, accelerate or change their paths.

[0083] 3. From 1 and 2, it can be known that the interaction between high-pressure and low-pressure systems is crucial for determining the movement of tropical cyclones.

[0084] In addition, based on the idea of local similarity, in the processing of the dataset, tropical cyclones will be truncated, on the one hand, to expand the number of datasets, and on the other hand, to focus on examining their local similarity.

[0085] Extract tropical cyclones that affect the waters of China or make landfall within 72 hours (the definition of affecting the waters of China is entering within the 72-hour warning line).

[0086] The same tropical cyclone is recalculated at 6-hour intervals for updating to generate a training dataset.

[0087] The process of generating and processing the training dataset is as follows.

[0088] ① For the CMA tropical cyclone best track dataset, process it one by one according to the tropical cyclone number, delete the sub-center situation (the processing method is to delete the records with sub-center identification), and form a path point dataset including tropical cyclone number, time, longitude, latitude, intensity, and central minimum pressure. The time interval between path points is 6h (hours).

[0089] ② For each tropical cyclone number (since 1949), extract from the two datasets of ERA5 hourly data on pressure levels from 1940 to present and ERA5 hourly data on single levels from 1940 to present according to its time (UTC), and extract four datasets (grid points) of 500hPa geopotential height, sea level pressure, and 10-meter wind (U, V) respectively; considering the spatial range matching with the CMA tropical cyclone path set (the northwest Pacific region), the spatial range of the above two ERA5 datasets is limited to a certain longitude and latitude range. The spatial range selected in this paper is [0 - 55°N, 90 - 180°E].

[0090] ③ For the 500hPa geopotential height grid data, extract the isopleth line greater than 5880gpm to obtain the subtropical high isosurface. The specific steps are as follows.

[0091] a) Based on the 500hPa geopotential height grid data, with 5880gpm as the condition, carry out isopleth line tracking and generation using the Marching Squares algorithm, and set the result as the first isopleth line result A.

[0092] b) Conduct a preliminary screening for A, and delete small-area closed curves (isopleth lines with less than 6 points), isolated points, etc. After the above processing, obtain the preliminarily screened isopleth line result B.

[0093] c) For B, use the Douglas–Peucker algorithm to simplify the generated isoclines. The preset threshold value is 3 - 4.5 times the shortest distance in B (i.e., the Euclidean distance between adjacent points in B). After being processed by the Douglas–Peucker algorithm, the simplified isocline result C is obtained, which is the isocline data of the 5880 gpm value.

[0094] d) For C, connect the points in the isoclines in sequence to generate the corresponding isosurface, that is, the subtropical high surface data, and this subtropical high surface data is the subtropical high isosurface.

[0095] ④ From the grid data of sea level pressure, 10 - meter wind U - component, and 10 - meter wind V - component, extract the tropical cyclone low - pressure area. At the same time, verify the low - pressure area through the strong - wind area (above level 6). The specific steps are as follows.

[0096] a) Based on the sea level pressure grid data, generate isoclines at intervals of 5 hPa from the preset 910 hPa to 1000 hPa. The isoclines are traced and generated using the Marching Squares algorithm, and the result is set as the second isocline result A1, which includes multiple isoclines.

[0097] b) For A1, perform screening processing item by item according to the threshold, and delete small - area closed curves (isoclines with less than 6 points), isolated points, etc. After the above - mentioned processing, the processed isocline result B1 is obtained, which includes multiple isoclines.

[0098] c) For B1, use the Douglas–Peucker algorithm to simplify the generated isoclines. The preset threshold value is 1.5 - 2.5 times the shortest distance in B1 (i.e., the Euclidean distance between adjacent points in B1). After being processed by the Douglas–Peucker algorithm, the simplified isocline result C is obtained.

[0099] d) Based on the grid data of 10 - meter wind U - component and 10 - meter wind V - component, generate gridded wind speed data , and its calculation formula is .

[0100] The threshold for strong winds above level 6 is set to 10.8 - 13.8 m / s, and the value taken in this application is 12 m / s.

[0101] According to the isocline generation method shown in ③, extract the isocline area of level 6 wind, and finally generate the isosurface.

[0102] e) According to the spatial range in step c) (select the low - pressure area of the 990 hPa isobar) and the coverage area in step d) (above level 6 wind), perform spatial intersection operation processing based on the GIS software to obtain the intersection range of the two.

[0103] f) After completing the above operations, the surface feature data of the low-pressure gale area is obtained, denoted as the low-pressure gale area D, as Figure 2 shown.

[0104] ⑤ Based on the space, the subtropical high isosurface and the low-pressure gale area D obtained in ③④ are subjected to fusion processing, that is, the two images are merged according to the corresponding longitude and latitude, and the merged result is E.

[0105] After the data processing is carried out through the above steps, the data set is organized in the following structure. Using the tropical cyclone number - time as the unique identification code, it is associated with the result E of step ⑤ at the corresponding time to establish a corresponding relationship, as Figure 3 shown.

[0106] In practical applications, the contrastive learning tropical cyclone similarity evaluation model can be a simple twin network (SimpleTwin Network Simsiam, Simsiam) model, and the present application also performs customized processing on the enhancement process of the Simsiam model.

[0107] The present application enhances the merged result E through the following two branches.

[0108] Branch one enhances the merged result E through the Canny algorithm. The Canny edge detection algorithm has four steps as follows: 1) Perform Gaussian filtering or median filtering on the image to filter noise; 2) Use the Sobel operator to calculate the gradient magnitude and direction of each pixel point in the image; 3) Use the non-maximum suppression algorithm to select the best edge among a group of edges. Specifically, check each pixel point and the pixel points with the same gradient direction nearby. If the gradient of the current pixel point is the largest, it is retained, otherwise it is removed; 4) Use double thresholds to determine the final edge. If the gradient of the pixel point is higher than the large threshold, it is retained; if the pixel point is lower than the small threshold, it is ignored; if it is between the two thresholds, determine whether the pixel point is connected to the edge pixel points.

[0109] After the above processing, the enhanced result X1 of the merged result E can be obtained.

[0110] Branch two intercepts a local image in the merged result E and performs binarization processing. The specific processing method is to take the center point of the tropical cyclone low pressure as the center point, intercept the image along the four directions of east, south, west, and north with a threshold of 5°, and at the same time perform binarization processing to generate the enhanced result X2.

[0111] The SimSiam model also includes an encoder, an MLP, a loss function, etc., which will not be elaborated here.

[0112] In addition, data from 1949 to 2023 are selected for model training. The training dataset is divided into a training set and a test set, with a ratio of 7:3.

[0113] This application not only considers the trajectory graph, but also considers the interaction relationship between the ocean and the atmosphere, comprehensively considers various factors to evaluate the similarity degree of tropical cyclones, and avoids misjudgment.

[0114] Based on the same inventive concept, an embodiment of this application also provides a tropical cyclone path screening system. The implementation solutions provided by this system to solve problems are similar to those described in the above method. Therefore, the specific limitations in one or more embodiments of the tropical cyclone path screening system provided below can refer to the limitations on the tropical cyclone path screening method in the above text, and will not be repeated here.

[0115] In an exemplary embodiment, a tropical cyclone path screening system is provided, which includes the following modules.

[0116] The first path similar typhoon list determination module is used to select a tropical cyclone sample and determine the first path similar typhoon list of the tropical cyclone sample based on the tropical cyclone path similarity method.

[0117] The second path similar typhoon list determination module is used to determine the second path similar typhoon list of the tropical cyclone sample based on the contrast learning tropical cyclone similarity evaluation model; the contrast learning tropical cyclone similarity evaluation model is trained by a training dataset; the training dataset includes the CMA tropical cyclone best track dataset and the ERA5 reanalysis data; the CMA tropical cyclone best track dataset includes the time, intensity, latitude, longitude, and central minimum pressure of the tropical cyclone; the ERA5 reanalysis data includes surface elements and meteorological elements at different pressure levels; the surface elements include sea level pressure and sea surface wind; the meteorological elements include geopotential height, wind, air temperature, vorticity, and divergence.

[0118] The best tropical cyclone path screening module is used to determine the tropical cyclone list similar to the path of the tropical cyclone sample according to the first path similar typhoon list and the second path similar typhoon list, and screen the best tropical cyclone path from the tropical cyclone list.

[0119] In an exemplary embodiment, a computer device is provided, which includes a memory and a processor. A computer program is stored in the memory. When the processor executes the computer program, the steps in the above method embodiments are implemented. This computer device can be a server or a terminal, and its internal structure diagram can be as Figure 4As shown in the figure. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data to be processed. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals through a network connection. When the computer program is executed by the processor, it implements a method for screening tropical cyclone paths.

[0120] Those skilled in the art can understand that Figure 4 the structure shown in the figure is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements. In an exemplary embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps in the above method embodiments are implemented.

[0121] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0122] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0123] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0124] Those of ordinary skill in the art can understand that all or part of the processes of implementing the methods in the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memories can include read-only memory (ROM), magnetic tapes, floppy disks, flash memories, optical memories, high-density embedded non-volatile memories, resistive random access memories (ReRAM), magnetoresistive random access memories (MRAM), ferroelectric random access memories (FRAM), phase change memories (PCM), graphene memories, etc. Volatile memories can include random access memory (RAM) or external cache memories, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0125] The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logics, data processing logics based on quantum computing, etc., without limitation.

[0126] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered to be within the scope described in this specification.

[0127] In this article, specific examples are used to elaborate on the principles and implementation manners of the present application. The description of the above embodiments is only used to help understand the method and its core idea of the present application; at the same time, for those of ordinary skill in the art, according to the idea of the present application, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present application.

Claims

1. A method for screening tropical cyclone paths, characterized in that, Including: Select tropical cyclone samples and determine the first list of typhoons with similar paths to the tropical cyclone samples based on the tropical cyclone path similarity method; Determine the second list of typhoons with similar paths to the tropical cyclone samples based on the contrast learning tropical cyclone similarity evaluation model; The contrast learning tropical cyclone similarity evaluation model is obtained by training with a training dataset; the training dataset includes the CMA tropical cyclone best track dataset and the ERA5 reanalysis data; the CMA tropical cyclone best track dataset includes the time, intensity, latitude, longitude, and central minimum pressure of tropical cyclones; the ERA5 reanalysis data includes surface elements and meteorological elements at different pressure levels; the surface elements include sea level pressure and sea surface wind; the meteorological elements include geopotential height, wind, air temperature, vorticity, and divergence; According to the first list of typhoons with similar paths and the second list of typhoons with similar paths, determine the list of tropical cyclones with similar paths to the tropical cyclone samples, and screen the best tropical cyclone path from the list of tropical cyclones.

2. The tropical cyclone path screening method according to claim 1, wherein Before determining the second list of typhoons with similar paths to the tropical cyclone samples based on the contrast learning tropical cyclone similarity evaluation model, it further includes: Process the training dataset to determine the processed training dataset.

3. The tropical cyclone path screening method according to claim 2, characterized in that Processing the training dataset to determine the processed training dataset specifically includes: Based on the tropical cyclone serial number, process each tropical cyclone in the CMA tropical cyclone best track dataset one by one to form a path point dataset; the processing includes deleting records with the sub - center identifier; Based on the tropical cyclone serial number, extract the 500hPa geopotential height grid data, sea level pressure data, 10 - meter wind U - component data, and 10 - meter wind V - component data from the ERA5 reanalysis data according to time; where U is the horizontal wind speed component; V is the vertical wind speed component; Determine the subtropical high isosurface according to the 500hPa geopotential height grid data; Determine the tropical cyclone low - pressure area according to the sea level pressure data, the 10 - meter wind U - component data, and the 10 - meter wind V - component data; Fuse the path point dataset, the subtropical high isosurface, and the tropical cyclone low - pressure area to construct the processed training dataset.

4. The tropical cyclone path screening method according to claim 3, characterized in that Determining the subtropical high isosurface according to the 500hPa geopotential height grid data specifically includes: Taking the isoline with a value greater than 5880gpm as the condition, perform isoline tracking on the 500hPa geopotential height grid data, and use the Marching Squares algorithm to determine the first isoline result; Perform a preliminary screening on the first isoline result to determine the preliminarily screened isoline result; Use the Douglas–Peucker algorithm to simplify the preliminarily screened isoline result to determine the simplified isoline result; Connect the points in the simplified isoline result in sequence to generate the subtropical high surface data; the subtropical high surface data is the subtropical high isosurface.

5. The tropical cyclone path screening method according to claim 3, wherein, Based on the sea level pressure data, the 10-meter wind U-component data, and the 10-meter wind V-component data, determine the tropical cyclone low-pressure area, specifically including: Taking the interval from 910 hPa to 1000 hPa and with an interval of 5 hPa, perform contour tracing on the sea level pressure data, and use the Marching Squares algorithm to determine the second contour result; Determine the low-pressure area of the 990 hPa isobar based on the second contour result; Generate gridded wind speed data based on the 10-meter wind U-component data and the 10-meter wind V-component data; Based on the gridded wind speed data, extract the contour area with wind speed above level 6; the contour area with wind speed above level 6 is the strong wind area with wind speed above level 6; Determine the low-pressure strong wind area based on the intersection area of the low-pressure area and the strong wind area; the low-pressure strong wind area is the tropical cyclone low-pressure area.

6. The tropical cyclone path screening method according to claim 5, wherein Determine the low-pressure area of the 990 hPa isobar based on the second contour result, specifically including: Process each item of the second contour result according to the threshold condition to determine the processed contour result; Use the Douglas–Peucker algorithm to simplify the processed contour result to determine the simplified contour result, and select the low-pressure area of the 990 hPa isobar according to the simplified contour result.

7. A tropical cyclone path screening system, characterized in that, Including: The first path-similar typhoon list determination module is used to select a tropical cyclone sample and determine the first path-similar typhoon list of the tropical cyclone sample based on the tropical cyclone path similarity method; The second path-similar typhoon list determination module is used to determine the second path-similar typhoon list of the tropical cyclone sample based on the contrastive learning tropical cyclone similarity evaluation model; The contrastive learning tropical cyclone similarity evaluation model is obtained by training with a training data set; the training data set includes the CMA tropical cyclone best track data set and the ERA5 reanalysis data; the CMA tropical cyclone best track data set includes the time, intensity, latitude, longitude, and central minimum pressure of the tropical cyclone; the ERA5 reanalysis data includes surface elements and meteorological elements at different pressure levels; the surface elements include sea level pressure and sea surface wind; the meteorological elements include geopotential height, wind, air temperature, vorticity, and divergence; The best tropical cyclone path screening module is used to determine the tropical cyclone list similar to the tropical cyclone sample path according to the first path-similar typhoon list and the second path-similar typhoon list, and screen the best tropical cyclone path from the tropical cyclone list.

8. A computer device, comprising: A memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor executes the computer program to implement the tropical cyclone path screening method according to any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the tropical cyclone path screening method according to any one of claims 1-6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the tropical cyclone path screening method according to any one of claims 1-6.

Images