Typhoon path integrated prediction method based on marine meteorological big data

Abstract

The invention discloses a typhoon path integrated prediction method based on marine meteorological big data, and the method comprises the steps: collecting multi-source marine meteorological data, carrying out the time and spatial resolution alignment and fusion of the multi-source marine meteorological data, and obtaining a marine meteorological model; extracting typhoon meteorological characteristics in the marine meteorological model, obtaining a historical typhoon data set, and predicting an initial action path of a typhoon according to the typhoon meteorological characteristics and the historical typhoon data set; detecting and acquiring an ocean temperature-salinity value on the initial action path, and correcting the initial action path according to the ocean temperature-salinity abnormal value to obtain a corrected action path; on the basis of a preset displacement prediction model, in combination with a typhoon passing path and the marine meteorological model, calculating a sudden change point and a turning probability of the typhoon; and according to the sudden change point and the steering probability thereof, correcting the corrected action path to obtain a final typhoon path, and displaying the final typhoon path on a geographic information platform.

Description

Technical Field

[0001] The present invention relates to the technical field of meteorological forecasting, and particularly to a typhoon path integrated prediction method based on marine meteorological big data. Background Art

[0002] Tropical cyclones (typhoons) are one of the most destructive natural disasters globally. Typhoons can form throughout the year in the northwestern Pacific Ocean where China is located, causing coastal and some inland areas to be frequently hit by typhoons. Therefore, typhoon research holds an important position in the meteorological community. Typhoon activities have significant statistical laws and historical similarity characteristics. Although modern satellite remote sensing and numerical forecasting technologies have been highly developed, the analysis of similar cases remains an important basis for typhoon forecasting and disaster prevention and mitigation decision-making. How to effectively utilize historical observation data, conduct typhoon similarity analysis, and provide more reliable similar reference information for forecasters, thereby improving the accuracy of typhoon forecasting, has important scientific and practical significance for disaster prevention and mitigation work.

[0003] Traditional typhoon path prediction mainly relies on numerical weather prediction models. Numerical weather prediction models predict typhoon paths by solving physical equations of atmospheric motion, but these models have certain limitations in resolution and the description of complex physical processes, resulting in difficulties in further improving prediction accuracy. Summary of the Invention

[0004] The present invention provides a typhoon path integrated prediction method based on marine meteorological big data to solve the technical problems in the prior art that the typhoon path prediction ability is weak, the resolution of existing models is low, resulting in high limitations in the description of typhoon processes, and high prediction accuracy cannot be achieved.

[0005] To solve the above technical problems, an embodiment of the present invention provides a typhoon path integrated prediction method based on marine meteorological big data, including: Collect multi-source marine meteorological data, and align and fuse the multi-source marine meteorological data in terms of time and space resolution to obtain a marine meteorological model; Extract typhoon meteorological characteristics from the marine meteorological model, and obtain a historical typhoon data set. According to the typhoon meteorological characteristics and the historical typhoon data set, predict the initial movement path of the typhoon; Detect and obtain the ocean temperature and salinity values on the initial movement path, and determine the ocean temperature and salinity anomaly values. According to the ocean temperature and salinity anomaly values, correct the initial movement path to obtain a corrected movement path; Based on a preset movement prediction model, combine the passing path of the typhoon and the marine meteorological model to calculate the mutation points of the typhoon and their turning probabilities; Based on the mutation points and their turning probabilities, correct the corrected action path to obtain the final typhoon path, and display the final typhoon path on the geographic information platform.

[0006] As a preferred solution, the collection of multi-source marine meteorological data and the alignment and fusion of the multi-source marine meteorological data in terms of time and space resolution to obtain a marine meteorological model specifically include: Collect multi-source marine meteorological data; the multi-source marine meteorological data includes: satellite remote sensing data, ground observation data, ocean current data, and radar data; Clean the collected satellite remote sensing data, ground observation data, ocean current data, and radar data respectively to remove the noise data and outliers in the satellite remote sensing data, ground observation data, ocean current data, and radar data; Based on a preset time resolution and a preset space resolution, uniformly align the satellite remote sensing data, ground observation data, ocean current data, and radar data to the spatio-temporal grid; Through the Kalman filtering algorithm, fuse the spatio-temporal grid and the satellite remote sensing data, ground observation data, ocean current data, and radar data therein to construct a marine meteorological model.

[0007] As a preferred solution, extract typhoon meteorological characteristics from the marine meteorological model, obtain a historical typhoon data set, and predict the initial action path of the typhoon according to the typhoon meteorological characteristics and the historical typhoon data set, specifically including: In the marine meteorological model, extract the atmospheric spatial distribution and the ocean circulation distribution, and extract the time series data corresponding to the atmospheric spatial distribution and the ocean circulation distribution in the marine meteorological model; wherein, the typhoon meteorological characteristics include: the atmospheric spatial distribution, the ocean circulation distribution, and the time characteristics corresponding to the atmospheric spatial distribution and the ocean circulation distribution respectively; Input the atmospheric spatial distribution, the ocean circulation distribution, and the corresponding time series data into a preset path prediction model, and output the future action path of the typhoon; Obtain a historical typhoon data set; wherein, the historical typhoon data set includes several historical typhoons and their corresponding historical paths; Based on the position of the current typhoon, determine the similarity between the path passed by the current typhoon and the historical path of the historical typhoon in the section before the position of the current typhoon, and extract the historical path corresponding to the historical typhoon with the highest similarity as the historical correction path; Preliminarily correct the future action path according to the historical correction path to obtain the initial action path.

[0008] As a preferred solution, the construction method of the preset path prediction model includes: Obtain historical typhoon sample data; the historical typhoon sample data includes: historical atmospheric spatial distribution samples, historical ocean circulation distribution samples, and historical path samples; Construct a deep learning model, and use the historical atmospheric spatial distribution samples, historical ocean circulation distribution samples, and time series data as inputs, and the historical path samples as outputs. Iteratively train the deep learning model until a preset number of iterations is reached to obtain a trained preset path prediction model.

[0009] As a preferred solution, detect and obtain the ocean temperature and salinity values on the initial action path, determine the ocean temperature and salinity anomaly values, and correct the initial action path according to the ocean temperature and salinity anomaly values to obtain a corrected action path. Specifically, it includes: Based on ocean buoy devices, detect the ocean temperature and salinity values at each position point on the initial action path; According to the current seasonal time, determine a preset ocean temperature and salinity range, and based on the ocean temperature and salinity range, determine the ocean temperature and salinity values outside the ocean temperature and salinity range as ocean temperature and salinity anomaly values, and use the position points belonging to the ocean temperature and salinity anomaly values as path correction points; where the number of the path correction points is at least 1; Correct the initial action path according to the path correction points and their corresponding ocean temperature and salinity anomaly values to obtain a corrected action path.

[0010] As a preferred solution, correct the initial action path according to the path correction points and their corresponding ocean temperature and salinity anomaly values to obtain a corrected action path. Specifically, it includes: During the process of correcting the initial action path at each path correction point, determine the first path correction point on the initial action path, and according to the ocean temperature and salinity anomaly value corresponding to the first path correction point, determine its correction direction and correction degree. Thus, according to the correction direction and correction degree corresponding to the first path correction point, correct the initial action path to obtain a corrected first action path. Then, starting from the corrected position point, re-determine the first path correction point on the corrected first action path, so as to re-correct the corrected first action path based on the re-determined first path correction point; until there are no ocean temperature and salinity anomaly values at each point on the corrected typhoon action path, the final corrected action path is obtained.

[0011] As a preferred solution, based on the preset position prediction model, combine the passing path of the typhoon and the ocean meteorological model to calculate the mutation points of the typhoon and their turning probabilities. Specifically, it includes: Construct a mathematical model based on a long short-term memory network as a preset movement prediction model; wherein, the preset movement prediction model is trained through historical typhoon paths and their historical time series data in a historical typhoon dataset; Input the passing path of the typhoon and the time series data of the corresponding passing path in the ocean meteorological model into the preset movement prediction model to obtain a movement action path; Compare the corrected action path with the movement action path to determine the mutation point of the typhoon, and based on the ocean meteorological model, calculate the turning direction and its turning probability of the typhoon at this mutation point.

[0012] Correspondingly, the present invention also provides a typhoon path integrated prediction system based on ocean meteorological big data, including: An alignment and fusion module for collecting multi-source ocean meteorological data, and aligning and fusing the multi-source ocean meteorological data in terms of time and space resolution to obtain an ocean meteorological model; A prediction module for extracting typhoon meteorological characteristics in the ocean meteorological model, obtaining a historical typhoon dataset, and predicting the initial action path of the typhoon according to the typhoon meteorological characteristics and the historical typhoon dataset; A correction module for detecting and obtaining the ocean temperature and salinity values on the initial action path, determining the ocean temperature and salinity anomaly values, and correcting the initial action path according to the ocean temperature and salinity anomaly values to obtain a corrected action path; A mutation module for calculating the mutation point and its turning probability of the typhoon based on a preset movement prediction model, in combination with the passing path of the typhoon and the ocean meteorological model; A deviation correction module for correcting the corrected action path according to the mutation point and its turning probability to obtain the final typhoon path, and displaying the final typhoon path on a geographic information platform.

[0013] Correspondingly, the present invention also provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the typhoon path integrated prediction method based on ocean meteorological big data as described in any one of the above.

[0014] Correspondingly, the present invention also provides a computer-readable storage medium, which includes a stored computer program. When the computer program runs, it controls the device where the computer-readable storage medium is located to execute the typhoon path integrated prediction method based on ocean meteorological big data as described in any one of the above.

[0015] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: By collecting multi-source marine meteorological data and aligning and fusing the multi-source marine meteorological data in time and space resolution, a marine meteorological model is constructed, thereby improving the temporal and spatial resolution of the mathematical model and avoiding the limitations of the typhoon process description. By extracting the typhoon meteorological characteristics and combining them with historical typhoon data sets, the typhoon's initial action path is predicted, so that the typhoon's action path can be predicted at the atmospheric distribution and ocean circulation levels. At the same time, the ocean temperature and salinity values ​​on the initial action path are obtained to correct the initial action path, increase the influence of ocean temperature and salinity on the predicted path, and improve the accuracy and dimension of the typhoon path prediction. Finally, the preset movement prediction model is combined to determine the mutation point of the corrected action path and calculate the turning probability, so as to correct the corrected action path and obtain the final typhoon path, thereby improving the ability to predict new and abnormal typhoon paths, so as to achieve a more accurate prediction of the typhoon path. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 : A flow chart of the steps of a typhoon path integrated prediction method based on marine meteorological big data provided by an embodiment of the present invention; Figure 2 : A structural diagram of a typhoon path integrated prediction system based on marine meteorological big data provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] Embodiment 1 Please refer to Figure 1 , a typhoon path integrated prediction method based on marine meteorological big data provided by an embodiment of the present invention, comprising the following steps S101-S105: Step S101: Collect multi-source marine meteorological data, and align and fuse the multi-source marine meteorological data in time and space resolution to obtain a marine meteorological model.

[0019] As a preferred solution of this embodiment, the multi-source marine meteorological data is collected, and the multi-source marine meteorological data is aligned and integrated in time and space resolution to obtain a marine meteorological model, specifically including: Collecting multi-source marine meteorological data; the multi-source marine meteorological data includes: satellite remote sensing data, ground observation data, ocean current data and radar data; Clean the collected satellite remote sensing data, ground observation data, ocean current data, and radar data respectively to remove the noise data and outliers in the satellite remote sensing data, ground observation data, ocean current data, and radar data; Based on the preset time resolution and preset space resolution, align the satellite remote sensing data, ground observation data, ocean current data, and radar data to the spatio-temporal grid uniformly; Through the Kalman filtering algorithm, fuse the spatio-temporal grid and the satellite remote sensing data, ground observation data, ocean current data, and radar data therein to construct an ocean meteorological model.

[0020] In this embodiment, the satellite remote sensing data includes information such as the position, intensity, cloud system structure, wind circle range, precipitation distribution, and atmospheric temperature and humidity profile of typhoons. The meteorological satellites equipped with instruments such as microwave scatterometers, microwave radiometers, infrared and visible light imagers can be used to collect data such as typhoon cloud images, wind fields, temperature and humidity. The ground observation data covers meteorological elements such as wind speed, wind direction, air pressure, precipitation, temperature, and humidity during the passage of typhoons, as well as ocean hydrological elements such as tide levels and waves. The ocean current data includes ocean current velocity and direction at different depths, sea surface current field distribution and other ocean current data, which can help analyze the impact of the ocean on typhoon intensity and path, and can be obtained through drifting buoys. The radar data is obtained through shipborne navigation radars installed on ships and offshore facilities, and can be used to monitor wave / swell height, period and direction, sea surface current velocity and direction, etc.

[0021] In this embodiment, clean the collected various types of data respectively to remove noise data and outliers, ensure data quality, avoid the interference of abnormal data on subsequent analysis, and provide a reliable data basis for the construction of the ocean meteorological model. Align the satellite remote sensing data, ground observation data, ocean current data, and radar data to the spatio-temporal grid uniformly based on the preset time resolution and space resolution, so that the data from different sources are consistent in the time and space dimensions, which is convenient for subsequent fusion processing. At the same time, it can ensure that the resolution of the model in time and space matches the actual needs, and improve the model's description and prediction ability of ocean meteorological phenomena.

[0022] In this embodiment, the Kalman filter algorithm is used to fuse the spatio-temporal grid and various types of data therein. The Kalman filter is a recursive algorithm based on the state space model. It can effectively fuse multi-source data and construct a marine meteorological model by dynamically estimating and updating the data. The Kalman filter is based on the state space model, which includes a state equation and an observation equation. The state equation describes the dynamic changes of multi-source marine meteorological data, while the observation equation relates the actual observed data (taking satellite remote sensing data, ground observation data, ocean current data, and radar data on the spatio-temporal grid as the input of the observed data) to the state variables. Then, using recursive estimation, that is, the Kalman filter, through a recursive manner, combines model prediction and actual observed data to perform an optimal estimation of the state variables, thereby performing state prediction, covariance prediction, Kalman gain calculation, state update, and covariance update. Finally, according to the reliability and correlation of the observed data, the weights of each data source are dynamically adjusted, giving priority to high-precision and high-reliability data, and thus realizing multi-source marine meteorological data. The fused marine meteorological model can more comprehensively reflect the state and change trend of the marine meteorological system, providing more accurate data support for marine meteorological forecasts such as typhoon path prediction.

[0023] Step S102: Extract typhoon meteorological features from the marine meteorological model, obtain a historical typhoon data set, and predict the initial action path of the typhoon according to the typhoon meteorological features and the historical typhoon data set.

[0024] As a preferred solution of this embodiment, extracting typhoon meteorological features from the marine meteorological model, obtaining a historical typhoon data set, and predicting the initial action path of the typhoon according to the typhoon meteorological features and the historical typhoon data set specifically include: In the marine meteorological model, extract the atmospheric spatial distribution and the ocean circulation distribution, and extract the time series data corresponding to the atmospheric spatial distribution and the ocean circulation distribution in the marine meteorological model; wherein, the typhoon meteorological features include: the atmospheric spatial distribution, the ocean circulation distribution, and the time features corresponding to the atmospheric spatial distribution and the ocean circulation distribution respectively; Input the atmospheric spatial distribution, the ocean circulation distribution, and the corresponding time series data into a preset path prediction model, and output the future action path of the typhoon; Obtain a historical typhoon data set; wherein, the historical typhoon data set includes a number of historical typhoons and their corresponding historical paths; Based on the position of the current typhoon, determine the similarity between the path that the current typhoon has passed through and the historical path of the historical typhoon in the section before the current typhoon position, and extract the historical path corresponding to the historical typhoon with the highest similarity as the historical correction path; Preliminarily correct the future action path according to the historical correction path to obtain the initial action path.

[0025] In this embodiment, the atmospheric spatial distribution and ocean circulation distribution data are extracted from the constructed ocean meteorological model. The atmospheric spatial distribution and ocean circulation distribution data can reflect the environmental conditions required for the occurrence and development of typhoons, such as atmospheric elements like temperature, humidity, and wind field, as well as ocean elements like ocean current velocity and direction. At the same time, the time series data corresponding to the atmospheric spatial distribution and ocean circulation distribution are extracted. Among them, the time series data can show the changes of each atmospheric spatial distribution and ocean circulation distribution over time during the development of typhoons, so as to construct a comprehensive typhoon meteorological feature set including atmospheric spatial distribution, ocean circulation distribution, and corresponding time characteristics.

[0026] In this embodiment, the above-extracted data are used as input features and input into a preset path prediction model. Among them, the preset path prediction model is constructed based on historical typhoon data and ocean meteorological knowledge, and can predict the possible future action path of a typhoon according to the input meteorological feature data.

[0027] In this embodiment, a historical typhoon data set containing a number of historical typhoons and their corresponding historical paths is collected, and based on the position of the current typhoon, the path passed by the current typhoon is compared with the historical path of the historical typhoon in the previous section before the current position. By calculating the similarity, the historical typhoon path closest to the path of the current typhoon is found, that is, the historical correction path. Taking the historical correction path as a reference, the future action path output by the prediction model is preliminarily corrected to obtain an initial action path, so that the historical observation data and the model prediction results can be combined to improve the accuracy and credibility of the predicted path.

[0028] In this embodiment, the historical correction path can preliminarily correct the future action path by calculating the prediction deviation between the historical correction path and the future action path obtained through the preset path prediction model under similar historical conditions, and applying this deviation to the current predicted path for correction. Further, the future action path output by the prediction model and the historical correction path can also be weighted and averaged based on the similarity to achieve the correction of the future action path.

[0029] In this embodiment, by integrating various key meteorological elements such as atmospheric spatial distribution and ocean circulation distribution and their time series data, more comprehensive and refined information is provided for typhoon path prediction, making the prediction result closer to the actual typhoon behavior. At the same time, by introducing the historical typhoon data set and extracting the historical correction path based on similarity calculation, the empirical value of historical data can be fully utilized to effectively correct the model prediction result, reduce the prediction error, and improve the reliability of the initial action path.

[0030] As a preferred solution of this embodiment, the construction method of the preset path prediction model includes: Obtain historical typhoon sample data; the historical typhoon sample data includes: historical atmospheric spatial distribution samples, historical ocean circulation distribution samples, and historical path samples; Construct a deep learning model, and use the historical atmospheric spatial distribution samples, historical ocean circulation distribution samples, and time series data as inputs, and the historical path samples as outputs. Iteratively train the deep learning model until a preset number of iterations is reached to obtain a trained preset path prediction model.

[0031] In this embodiment, historical typhoon sample data is collected, including historical atmospheric spatial distribution samples (such as temperature, humidity, wind field, etc.), historical ocean circulation distribution samples (such as ocean current velocity, direction, etc.), and historical path samples (such as typhoon longitude and latitude, time, etc.). Among them, the historical typhoon sample data can be obtained from multiple sources such as meteorological satellites, ground observation stations, and ocean buoys. Furthermore, a deep learning model suitable for time series prediction is selected, such as Long Short-Term Memory (LSTM), Convolutional Neural Network (CNN) combined with attention mechanism, etc. Among them, LSTM is suitable for processing sequence data and can capture time-dependent relationships; CNN combined with attention mechanism can extract spatial features and focus on key information. Use the historical atmospheric spatial distribution samples, historical ocean circulation distribution samples, and time series data as model inputs, and the historical path samples as outputs to construct a neural network including an input layer, a hidden layer, and an output layer. The hidden layer can use an LSTM layer or a CNN layer combined with an attention mechanism to extract time series and spatial features. Divide the processed data set into a training set and a validation set, use the training set to iteratively train the model, calculate the predicted value through forward propagation, use a loss function (such as Mean Squared Error, MSE) to evaluate the difference between the predicted value and the true value, and then update the model parameters through backpropagation and an optimization algorithm (such as Adam) to reduce the loss. At the same time, adjust hyperparameters such as the learning rate, batch size, and number of iterations to optimize the model performance. Finally, obtain a trained preset path prediction model.

[0032] Step S103: Detect and obtain the ocean temperature and salinity values on the initial action path, determine the ocean temperature and salinity outliers, and correct the initial action path according to the ocean temperature and salinity outliers to obtain a corrected action path.

[0033] As a preferred solution of this embodiment, the detecting and obtaining the ocean temperature and salinity values on the initial action path, determining the ocean temperature and salinity outliers, and correcting the initial action path according to the ocean temperature and salinity outliers to obtain a corrected action path specifically includes: Based on ocean buoy devices, detect the ocean temperature and salinity values at each position point on the initial action path; Determine a preset ocean temperature and salinity range according to the current seasonal time, and based on the ocean temperature and salinity range, determine the ocean temperature and salinity values outside the ocean temperature and salinity range as ocean temperature and salinity anomaly values, and use the position points belonging to the ocean temperature and salinity anomaly values as path correction points; wherein, the number of the path correction points is at least 1; Correct the initial action path according to the path correction points and their corresponding ocean temperature and salinity anomaly values to obtain a corrected action path.

[0034] In this embodiment, first, based on the ocean buoy device, detect the ocean temperature and salinity values of each position point on the first action path, then determine the preset ocean temperature and salinity range according to the current season, and set the values exceeding this range as ocean temperature and salinity anomaly values, and the corresponding position points are the path correction points. Finally, adjust the initial action path according to these correction points and their temperature and salinity anomaly values to obtain the corrected action path.

[0035] In this embodiment, the ocean temperature and salinity detection device can be an ocean buoy device, which is equipped with a temperature sensor (such as a platinum resistance) and a conductivity sensor. The platinum resistance temperature measurement is based on the principle that the resistance changes linearly with temperature, and the conductivity sensor detects the conductivity of seawater to calculate the salinity. Some buoys are also equipped with a salinometer to directly measure the salinity. After the buoy enters the water, the sensors start to measure the temperature and conductivity of the seawater. The measured data can be transmitted to the onshore data center through satellite communication, etc. after being sorted out by the data acquisition system. Some buoys have the ability to process data in real time and can perform simple smoothing and correction processing.

[0036] In this embodiment, according to the current seasonal time, refer to the historical ocean temperature and salinity data and oceanographic research to determine a reasonable preset ocean temperature and salinity range. Among them, this range will change in different sea areas and seasons and can be set according to the actual situation. Compare the detected ocean temperature and salinity values with the preset range, and those exceeding the range are the ocean temperature and salinity anomaly values. Among them, the path points corresponding to the anomaly values are the path correction points, and the number is at least one. Exemplarily, based on the probability statistical characteristics of ocean observation data in the region, test whether the temperature and salinity elements obey the seasonal change statistical law. The annual / month average value a of the statistical station position element data can be passed mn and the corresponding mean square deviation sd are selected, and a reasonable multiple m (preferably, m = 3 can be selected). If the observed value X does not satisfy the formula: a mn - m×sd ≤ X ≤ a mn + m×sd, then it is determined as an anomaly value and needs further analysis.

[0037] In this embodiment, according to the path correction point and its ocean temperature and salinity anomaly value, a correction model or algorithm, such as an interpolation method, a statistical regression model or a physical ocean model, is established to consider the impact mechanism of ocean temperature and salinity anomaly on the typhoon path, such as changing the sea-air heat and water vapor exchange, ocean circulation, etc., and adjust the initial action path. The anomaly of ocean temperature and salinity is applied to the initial action path to obtain a corrected action path, which is more in line with the actual marine environmental conditions. At the same time, the buoy equipment is used to directly detect the ocean temperature and salinity value, provide rich measured data, avoid data loss or rely on indirect calculations of other data sources, and improve the scientificity and reliability of path correction. Changes in the ocean temperature and salinity structure can affect the heat transfer from the ocean to the atmosphere. For example, abnormal changes in the sea surface temperature may affect the intensity and movement speed of the typhoon. Warmer sea surface temperatures can provide more heat and water vapor for the typhoon, thereby enhancing the intensity of the typhoon, and the ocean temperature and salinity anomaly may cause changes in the ocean circulation, thereby affecting the path of the typhoon. Therefore, this embodiment can also avoid the actual situation of only considering the influence of atmospheric distribution and ocean circulation in the prior art, while ignoring the influence of ocean temperature and salinity on the typhoon.

[0038] As a preferred solution of this embodiment, the initial action path is corrected according to the path correction point and its corresponding ocean temperature and salinity anomaly value to obtain a corrected action path, specifically including: In the process of correcting the initial action path at each path correction point, the first path correction point on the initial action path is determined, and the correction direction and degree are determined according to the ocean temperature and salinity anomaly corresponding to the first path correction point, so that the initial action path is corrected according to the correction direction and degree corresponding to the first path correction point to obtain a corrected first action path, and then starting from the corrected position point, the first path correction point is re-determined on the corrected first action path, so that the corrected first action path is re-corrected based on the re-determined first path correction point; until there are no ocean temperature and salinity anomalies at each point on the corrected typhoon action path, the final corrected action path is obtained.

[0039] In this embodiment, on the initial action path, starting from the current position point of the typhoon as the starting point, each position point is checked in sequence to find the first point where the ocean temperature and salinity value exceeds the preset range, which is the first path correction point. Furthermore, the correction direction and degree are determined, that is, based on the ocean temperature and salinity anomaly value of the first path correction point, combined with oceanographic principles and the typhoon path prediction model, the correction direction (such as shifting to the left or right) and the correction degree (the distance or angle of the shift) are determined. For example, if the ocean temperature and salinity anomaly value causes the typhoon to possibly shift in a certain direction. Then, through the determined correction direction and degree, the initial action path is adjusted to obtain the first corrected action path. Then, the path correction point is re-determined. Starting from the path point after the last correction, the subsequent position points are continuously checked on the first corrected action path, the first path correction point on this path is re-determined, the correction direction and degree are determined again according to the ocean temperature and salinity anomaly value corresponding to the new path correction point, and the path is corrected. This process is repeated until there are no ocean temperature and salinity anomalies at each point on the corrected typhoon action path.

[0040] In this embodiment, when the ocean temperature and salinity values of all position points on the corrected typhoon action path are within the preset range, the correction process is terminated to obtain the final corrected action path, so as to be able to monitor the ocean temperature and salinity anomalies in real time and dynamically adjust the typhoon path prediction according to the anomaly situation, thereby more accurately reflecting the actual path change of the typhoon. At the same time, by analyzing the ocean temperature and salinity anomaly values of each path correction point in sequence and re-determining the specific correction direction and degree, the path correction is made more targeted and effective. Finally, through the iterative correction process, the influence of ocean temperature and salinity anomalies on the path is gradually eliminated, and a relatively accurate typhoon action path is finally obtained.

[0041] Step S104: Based on the preset movement prediction model, combined with the passing path of the typhoon and the ocean meteorological model, calculate the mutation point of the typhoon and its turning probability.

[0042] As a preferred solution of this embodiment, the calculating the mutation point of the typhoon and its turning probability based on the preset movement prediction model, combined with the passing path of the typhoon and the ocean meteorological model, specifically includes: Construct a mathematical model based on the long short-term memory network as the preset movement prediction model; wherein, the preset movement prediction model is trained by the historical typhoon paths and their historical time series data in the historical typhoon dataset; Input the passing path of the typhoon and the time series data of the corresponding passing path in the ocean meteorological model into the preset movement prediction model to obtain the movement action path; Compare the corrected action path and the movement action path to determine the mutation point of the typhoon, and based on the ocean meteorological model, calculate the turning direction and the turning probability of the typhoon at this mutation point.

[0043] In this embodiment, an LSTM neural network model including an input layer, a hidden layer (LSTM layer), and an output layer is constructed. The input layer receives the processed historical typhoon paths and their time series data. The hidden layer uses the LSTM structure to extract time-dependent features, and the output layer predicts the future typhoon paths. A historical typhoon data set including historical typhoon paths and corresponding time series data (such as atmospheric and ocean data) is collected. After data preprocessing (cleaning, standardization) and feature engineering (extracting time steps, constructing supervised learning problems), it is divided into a training set and a validation set. Then, the training set is used to iteratively train the LSTM model, and the model parameters are updated through an optimization algorithm (such as Adam) and a loss function (such as mean squared error) to minimize the difference between the predicted value and the true value and improve the prediction accuracy. Among them, the training continues until a preset number of iterations is reached or the convergence condition is satisfied.

[0044] In this embodiment, the passing path of the typhoon and the corresponding time series data are extracted from the ocean meteorological model and preprocessed. The format needs to be the same as the input during model training. The processed data is input into the trained LSTM model to obtain the predicted movement action path. Furthermore, the corrected action path and the movement action path are compared point by point, the differences at each point are calculated and the mutation points are identified, and a threshold is set. When the path difference exceeds the limit, this point is determined as a mutation point. Among them, the threshold can be set according to the actual situation.

[0045] In this embodiment, based on the ocean meteorological model, the ocean environmental characteristics (such as temperature-salinity distribution, ocean current) and atmospheric conditions (such as wind field, air pressure) at the mutation point are analyzed to predict the turning direction of the typhoon, and according to the historical typhoon data and model analysis, the probability of the typhoon turning at the mutation point is calculated to provide a quantitative assessment.

[0046] In this embodiment, to calculate the direction of the typhoon's turning and its turning probability, the longitude and latitude of two position points before and after the mutation point can be converted into radian values, and the moving direction of the typhoon at the mutation point can be calculated using the longitude and latitude radian values before and after the mutation point. That is, by calculating the change in longitude and latitude of two position points before and after the mutation point, the moving direction of the typhoon can be obtained using spherical trigonometry formulas. This method takes into account the influence of the earth's curvature and can more accurately reflect the moving direction of the typhoon in the spherical coordinate system. The turning probability can be calculated by collecting historical typhoon data, including the turning situations of typhoons under different conditions, especially the turning directions and frequencies when the environment is similar to the current mutation point (such as similar ocean temperature and salinity conditions, atmospheric circulation patterns, etc.). The number of times the typhoon turns in each direction under similar conditions is statistically counted to calculate the turning probability. For example, if in 100 historical cases similar to the current mutation point environment, the typhoon turns left 70 times and turns right 30 times, then the probability of turning left is 70%, and the probability of turning right is 30%.

[0047] It can be understood that the method based on historical data statistics utilizes the similarity and historical repeatability principles of typhoon paths and assumes that under similar environmental conditions, the turning behavior of typhoons has certain statistical regularity.

[0048] In this embodiment, the typhoon path prediction can also be corrected by how to use the Dynamic Time Warping (DTW) algorithm. Among them, the Dynamic Time Warping (DTW) algorithm is a method for measuring the similarity of two time series, especially suitable for time series with inconsistent lengths. In typhoon path prediction, the DTW algorithm can correct the prediction results by comparing the predicted path with historical similar paths. The core idea of the DTW algorithm is to find the optimal matching typhoon path between two time series through dynamic programming, so that the cumulative distance on the path is minimized.

[0049] First, construct a distance matrix: Assume that the two time series are Q=(q1,q2,…,qn) and C=(c1,c2,…,cm), with lengths n and m respectively. Construct an n×m matrix D, where D[i,j] represents the distance between qi and cj (usually using the Euclidean distance).

[0050] Cumulative distance calculation: Define the cumulative distance matrix D. Starting from D[1,1], calculate the cumulative distance of each point step by step. The calculation formula for the cumulative distance is: D[i,j]=d(qi,cj)+min{D[i - 1,j],D[i,j - 1],D[i - 1,j - 1]} Among them, d(qi,cj) is the distance between qi and cj.

[0051] Finding the optimal typhoon path: Starting from the lower right corner D[n,m] of the matrix, trace back to the upper left corner D[1,1] to find the typhoon path with the minimum cumulative distance, which is the optimal matching path.

[0052] Furthermore, set the constraint conditions: The path must start from the lower left corner (1,1) and end at the upper right corner (n,m). At the same time, ensure continuity and monotonicity. Among them, continuity means that the next point of the path can only be the point on the right, below, or lower right of the current point; monotonicity means that the path must increase monotonically with time and no backtracking is allowed.

[0053] In this embodiment, the DTW algorithm can be used to compare the predicted path with the historical similar paths for typhoon path prediction and correct the prediction results: Collect historical typhoon path data and current predicted path data, and convert them into time series format. Use the DTW algorithm to calculate the similarity between the current predicted path and the historical paths, and find the most similar historical path. Adjust the current predicted path according to the most similar historical path. For example, if there is an obvious deviation in a certain area of the historical path, the predicted path can be adjusted in that direction.

[0054] Furthermore, adjust the predicted path according to the optimal matching path to make it closer to the historical path. For example, if there is a large deviation between the predicted path and the historical path at a certain point, it can be corrected by interpolation or smoothing.

[0055] Step S105: Correct the corrected action path according to the mutation point and its turning probability to obtain the final typhoon path, and display the final typhoon path on the geographic information platform.

[0056] In this embodiment, by using historical typhoon data and the characteristics of the current typhoon, calculate the turning probability of the typhoon at each mutation point, analyze the change in the moving direction of the typhoon at the mutation point, determine the possible turning direction, and then correct each mutation point on the initial action path according to the turning probability and turning direction. For the direction with a higher turning probability, appropriately adjust the offset of the path to make the path closer to the possible actual trend. Repeat the above steps, continuously update the current position and path prediction of the typhoon until the typhoon ends or reaches the predetermined prediction duration.

[0057] In this embodiment, the turning probability reflects the possibility of a typhoon turning in different directions at the mutation point. By considering these probabilities in path correction, the corrected path can be made more consistent with the actual behavior pattern of the typhoon, enabling this embodiment to combine various factors in the ocean meteorological model (such as temperature-salinity distribution, ocean currents, atmospheric circulation, etc.) to more comprehensively evaluate the possibility of typhoon turning, improve the accuracy of path correction, and continuously update the model input and prediction results with the movement of the typhoon and the acquisition of new data, and timely adjust the path correction strategy to ensure the accuracy and timeliness of the final path.

[0058] Furthermore, finally, the final typhoon path is displayed on the geographic information platform, enabling users to intuitively understand the typhoon's movement direction and the manifestation of future predicted results, thus improving the user experience.

[0059] Implementing the above embodiments has the following effects: By collecting multi-source ocean meteorological data and aligning and fusing the multi-source ocean meteorological data in terms of time and space resolution to construct an ocean meteorological model, the resolution of the mathematical model in time and space is improved, avoiding the limitations of typhoon process description. By extracting typhoon meteorological characteristics from it and combining with historical typhoon data sets, the initial action path of the typhoon is predicted, enabling the prediction of the typhoon's action path at the atmospheric distribution and ocean circulation levels. At the same time, the ocean temperature-salinity values on the initial action path are obtained to correct the initial action path, realizing the addition of the influence of ocean temperature-salinity on the predicted path, improving the accuracy and dimension of typhoon path prediction. Finally, combined with a preset movement prediction model, the mutation points of the corrected action path are determined and the turning probabilities are calculated to correct the corrected action path, obtaining the final typhoon path, improving the prediction ability for new and abnormal typhoon paths, and achieving more accurate prediction of typhoon paths.

[0060] Embodiment Two Please refer to Figure 2 , which is a typhoon path integrated prediction system provided by the present invention based on ocean meteorological big data, including: An alignment and fusion module 201, configured to collect multi-source ocean meteorological data and align and fuse the multi-source ocean meteorological data in terms of time and space resolution to obtain an ocean meteorological model; A prediction module 202, configured to extract typhoon meteorological characteristics from the ocean meteorological model, obtain a historical typhoon data set, and predict the initial action path of the typhoon according to the typhoon meteorological characteristics and the historical typhoon data set; A correction module 203, configured to detect and obtain the ocean temperature-salinity values on the initial action path, determine the ocean temperature-salinity anomaly values, and correct the initial action path according to the ocean temperature-salinity anomaly values to obtain a corrected action path; A mutation module 204, configured to calculate the mutation points of the typhoon and their turning probabilities based on a preset movement prediction model, in combination with the passing path of the typhoon and the ocean meteorological model. A deviation correction module 205, configured to correct the corrected action path according to the mutation points and their turning probabilities to obtain the final typhoon path, and display the final typhoon path on a geographic information platform.

[0061] As a preferred solution, collecting multi-source ocean meteorological data, and aligning and fusing the multi-source ocean meteorological data in terms of time and space resolution to obtain an ocean meteorological model, specifically including: Collecting multi-source ocean meteorological data; the multi-source ocean meteorological data includes: satellite remote sensing data, ground observation data, ocean current data, and radar data; Cleaning the collected satellite remote sensing data, ground observation data, ocean current data, and radar data respectively to remove the noise data and outliers in the satellite remote sensing data, ground observation data, ocean current data, and radar data; Based on a preset time resolution and a preset space resolution, uniformly aligning the satellite remote sensing data, ground observation data, ocean current data, and radar data to a spatio-temporal grid; Through a Kalman filtering algorithm, fusing the spatio-temporal grid and the satellite remote sensing data, ground observation data, ocean current data, and radar data therein to construct an ocean meteorological model.

[0062] As a preferred solution, extracting typhoon meteorological features from the ocean meteorological model, and obtaining a historical typhoon data set. According to the typhoon meteorological features and the historical typhoon data set, predicting the initial action path of the typhoon, specifically including: In the ocean meteorological model, extracting the atmospheric spatial distribution and the ocean circulation distribution, and extracting the time series data corresponding to the atmospheric spatial distribution and the ocean circulation distribution in the ocean meteorological model; wherein, the typhoon meteorological features include: the atmospheric spatial distribution, the ocean circulation distribution, and the time features corresponding to the atmospheric spatial distribution and the ocean circulation distribution respectively; Inputting the atmospheric spatial distribution, the ocean circulation distribution, and the corresponding time series data respectively into a preset path prediction model, and outputting the future action path of the typhoon; Obtaining a historical typhoon data set; wherein, the historical typhoon data set includes a number of historical typhoons and their corresponding historical paths; Based on the position of the current typhoon, determining the similarity between the path passed by the current typhoon and the historical path of the historical typhoon in the section before the current typhoon position, and extracting the historical path corresponding to the historical typhoon with the highest similarity as the historical correction path; Perform a preliminary correction on the future action path according to the historical correction path to obtain an initial action path.

[0063] As a preferred solution, the method for constructing the preset path prediction model includes: Obtain historical typhoon sample data; the historical typhoon sample data includes: historical atmospheric spatial distribution samples, historical ocean circulation distribution samples, and historical path samples; Construct a deep learning model, and use the historical atmospheric spatial distribution samples, historical ocean circulation distribution samples, and time series data as inputs, and the historical path samples as outputs. Iteratively train the deep learning model until a preset number of iterations is reached to obtain a trained preset path prediction model.

[0064] As a preferred solution, detect and obtain the ocean temperature and salinity values on the initial action path, determine the ocean temperature and salinity anomaly values, and correct the initial action path according to the ocean temperature and salinity anomaly values to obtain a corrected action path. Specifically, it includes: Based on ocean buoy devices, detect the ocean temperature and salinity values at each position point on the initial action path; According to the current seasonal time, determine a preset ocean temperature and salinity range, and based on the ocean temperature and salinity range, determine the ocean temperature and salinity values outside the ocean temperature and salinity range as ocean temperature and salinity anomaly values, and use the position points belonging to the ocean temperature and salinity anomaly values as path correction points; where the number of the path correction points is at least 1; Correct the initial action path according to the path correction points and their corresponding ocean temperature and salinity anomaly values to obtain a corrected action path.

[0065] As a preferred solution, correct the initial action path according to the path correction points and their corresponding ocean temperature and salinity anomaly values to obtain a corrected action path. Specifically, it includes: During the process of correcting the initial action path at each path correction point, determine the first path correction point on the initial action path, and according to the ocean temperature and salinity anomaly value corresponding to the first path correction point, determine its correction direction and correction degree. Then, according to the correction direction and correction degree corresponding to the first path correction point, correct the initial action path to obtain a corrected first action path. Furthermore, starting from the corrected position point, re-determine the first path correction point on the corrected first action path, so that based on the re-determined first path correction point, the corrected first action path is corrected again; until there are no ocean temperature and salinity anomaly values at each point on the corrected typhoon action path, the final corrected action path is obtained.

[0066] As a preferred solution, based on the preset movement prediction model, combining the passing path of the typhoon and the ocean meteorological model, the mutation point of the typhoon and its turning probability are calculated, specifically including: Construct a mathematical model based on the long short-term memory network as the preset movement prediction model; wherein, the preset movement prediction model is trained by the historical typhoon paths and their historical time series data in the historical typhoon dataset; Input the passing path of the typhoon and the time series data of the corresponding passing path in the ocean meteorological model into the preset movement prediction model to obtain the movement action path; Compare the corrected action path and the movement action path to determine the mutation point of the typhoon, and based on the ocean meteorological model, calculate the turning direction and its turning probability of the typhoon at this mutation point.

[0067] Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working process of the above-described device can refer to the corresponding process in the foregoing method embodiment and will not be elaborated herein.

[0068] Implementing the above embodiments has the following effects: By collecting multi-source ocean meteorological data, aligning and fusing the multi-source ocean meteorological data in terms of time and space resolution to construct an ocean meteorological model, thereby improving the resolution of the mathematical model in time and space, avoiding the limitations of typhoon process description, and by extracting the typhoon meteorological characteristics therein and combining with the historical typhoon dataset to predict the initial action path of the typhoon, so as to be able to predict the action path of the typhoon at the level of atmospheric distribution and ocean circulation. At the same time, obtain the ocean temperature and salinity values on the initial action path to correct the initial action path, realizing the addition of the influence of ocean temperature and salinity on the basis of the predicted path, improving the accuracy and dimension of typhoon path prediction, and finally combining the preset movement prediction model to determine the mutation point and calculate the turning probability of the corrected action path, so as to correct the corrected action path to obtain the final typhoon path, improving the prediction ability for new and abnormal typhoon paths, so as to achieve more accurate prediction of typhoon paths.

[0069] Embodiment III Correspondingly, the present invention also provides a terminal device, including: a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the typhoon path integrated prediction method based on ocean meteorological big data as described in any one of the above embodiments.

[0070] The terminal device of this embodiment includes: a processor, a memory, and a computer program and computer instructions stored in the memory and executable on the processor. When the processor executes the computer program, it implements each step in the first embodiment above, such as Figure 1 the steps S101 to S105 shown. Alternatively, when the processor executes the computer program, it implements the functions of each module / unit in the above device embodiment, such as the deviation correction module 205.

[0071] Exemplarily, the computer program can be divided into one or more modules / units. The one or more modules / units are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units can be a series of computer program instruction segments capable of performing specific functions, and this instruction segment is used to describe the execution process of the computer program in the terminal device. For example, the deviation correction module 205 is used to correct the modified action path according to the mutation point and its turning probability to obtain the final typhoon path, and display the final typhoon path on the geographic information platform.

[0072] The terminal device can be a computing device such as a desktop computer, a notebook, a palm computer, and a cloud server. The terminal device may include, but is not limited to, a processor and a memory. Those skilled in the art can understand that the schematic diagram is only an example of the terminal device and does not constitute a limitation on the terminal device. It may include more or fewer components than shown in the figure, or combine some components, or different components. For example, the terminal device may further include input / output devices, network access devices, a bus, etc.

[0073] The so-called processor may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), off-the-shelf programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The processor is the control center of the terminal device, and connects various parts of the entire terminal device through various interfaces and lines.

[0074] The memory can be used to store the computer programs and / or modules. By running or executing the computer programs and / or modules stored in the memory and invoking the data stored in the memory, the processor realizes various functions of the terminal device. The memory mainly includes a program storage area and a data storage area. Among them, the program storage area can store an operating system, application programs required for at least one function, etc.; the data storage area can store data created according to the use of the mobile terminal, etc. In addition, the memory can include high-speed random access memory, and can also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one magnetic disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0075] Among them, if the module / unit integrated in the terminal device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, to implement all or part of the processes in the above-mentioned embodiment methods of the present invention, it can also be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file, or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disc, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc. It should be noted that the content included in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0076] Embodiment 4 Correspondingly, the present invention also provides a computer-readable storage medium. The computer-readable storage medium includes a stored computer program. Among them, when the computer program runs, it controls the device where the computer-readable storage medium is located to execute the typhoon path integrated prediction method based on ocean meteorological big data described in any one of the above embodiments.

[0077] The specific embodiments described above further elaborate on the objectives, technical solutions, and beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the protection scope of the present invention. In particular, for those skilled in the art, any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A typhoon path integrated prediction method based on marine meteorological big data, characterized in that, Including: Collect multi-source marine meteorological data, align and fuse the multi-source marine meteorological data in terms of time and space resolution to obtain a marine meteorological model; Extract typhoon meteorological characteristics from the marine meteorological model, obtain a historical typhoon data set, and predict the initial action path of the typhoon according to the typhoon meteorological characteristics and the historical typhoon data set; Detect and obtain the ocean temperature and salinity values on the initial action path, determine the ocean temperature and salinity anomaly values, and correct the initial action path according to the ocean temperature and salinity anomaly values to obtain a corrected action path; Based on a preset movement prediction model, combine the passing path of the typhoon and the marine meteorological model to calculate the mutation points of the typhoon and their turning probabilities; Correct the corrected action path according to the mutation points and their turning probabilities to obtain the final typhoon path, and display the final typhoon path on a geographic information platform.

2. The typhoon path integrated prediction method based on marine meteorological big data according to claim 1, wherein, The step of collecting multi-source marine meteorological data, aligning and fusing the multi-source marine meteorological data in terms of time and space resolution to obtain a marine meteorological model specifically includes: Collect multi-source marine meteorological data; the multi-source marine meteorological data includes: satellite remote sensing data, ground observation data, ocean current data, and radar data; Clean the collected satellite remote sensing data, ground observation data, ocean current data, and radar data respectively to remove the noise data and outliers in the satellite remote sensing data, ground observation data, ocean current data, and radar data; Based on a preset time resolution and a preset space resolution, uniformly align the satellite remote sensing data, ground observation data, ocean current data, and radar data to a spatio-temporal grid; Through the Kalman filter algorithm, fuse the spatio-temporal grid and the satellite remote sensing data, ground observation data, ocean current data, and radar data therein to construct a marine meteorological model.

3. The typhoon path integrated prediction method based on marine meteorological big data according to claim 2, wherein, The step of extracting typhoon meteorological characteristics from the marine meteorological model, obtaining a historical typhoon data set, and predicting the initial action path of the typhoon according to the typhoon meteorological characteristics and the historical typhoon data set specifically includes: In the marine meteorological model, extract the atmospheric spatial distribution and the ocean circulation distribution, and extract the time series data corresponding to the atmospheric spatial distribution and the ocean circulation distribution in the marine meteorological model; wherein, the typhoon meteorological characteristics include: the atmospheric spatial distribution, the ocean circulation distribution, and the time characteristics corresponding to the atmospheric spatial distribution and the ocean circulation distribution respectively; Input the atmospheric spatial distribution, the ocean circulation distribution, and the corresponding time series data into a preset path prediction model, and output the future action path of the typhoon; Obtain a historical typhoon data set; wherein, the historical typhoon data set includes a number of historical typhoons and their corresponding historical paths; Based on the position of the current typhoon, determine the similarity between the path passed by the current typhoon and the historical path of the historical typhoon in the segment before the current typhoon position, and extract the historical path corresponding to the historical typhoon with the highest similarity as the historical correction path; Preliminarily correct the future action path according to the historical correction path to obtain an initial action path.

4. The typhoon track integrated prediction method based on marine meteorological big data according to claim 3, characterized in that, The construction method of the preset path prediction model includes: Obtain historical typhoon sample data; the historical typhoon sample data includes: historical atmospheric spatial distribution samples, historical ocean circulation distribution samples, and historical path samples; Construct a deep learning model, and use the historical atmospheric spatial distribution samples, historical ocean circulation distribution samples, and time series data as inputs, and the historical path samples as outputs, and perform iterative training on the deep learning model until a preset number of iterations is reached to obtain a trained preset path prediction model.

5. The typhoon path integrated prediction method based on marine meteorological big data according to claim 1, characterized in that, Detect and obtain the ocean temperature and salinity values on the initial action path, and determine the ocean temperature and salinity anomaly values, and correct the initial action path according to the ocean temperature and salinity anomaly values to obtain a corrected action path, which specifically includes: Based on ocean buoy devices, detect the ocean temperature and salinity values at each position point on the initial action path; According to the current seasonal time, determine a preset ocean temperature and salinity range, and based on the ocean temperature and salinity range, determine the ocean temperature and salinity values outside the ocean temperature and salinity range as ocean temperature and salinity anomaly values, and use the position points belonging to the ocean temperature and salinity anomaly values as path correction points; where the number of the path correction points is at least 1; Correct the initial action path according to the path correction points and their corresponding ocean temperature and salinity anomaly values to obtain a corrected action path.

6. The typhoon path integrated prediction method based on marine meteorological big data according to claim 5, characterized in that Correct the initial action path according to the path correction points and their corresponding ocean temperature and salinity anomaly values to obtain a corrected action path, which specifically includes: During the process of correcting the initial action path at each path correction point, determine the first path correction point on the initial action path, and according to the ocean temperature and salinity anomaly value corresponding to the first path correction point, determine its correction direction and correction degree, so as to correct the initial action path according to the correction direction and correction degree corresponding to the first path correction point to obtain a corrected first action path, and then start from the corrected position point, re-determine the first path correction point on the corrected first action path, so that based on the re-determined first path correction point, the corrected first action path is corrected again; until there are no ocean temperature and salinity anomaly values at each point on the corrected typhoon action path, the final corrected action path is obtained.

7. A typhoon track integrated prediction method based on marine meteorological big data according to any one of claims 1-6, characterized in that Based on the preset movement prediction model, combine the passing path of the typhoon and the ocean meteorological model to calculate the mutation point of the typhoon and its turning probability, which specifically includes: Construct a mathematical model based on a long short-term memory network as the preset movement prediction model; where the preset movement prediction model is trained by the historical typhoon paths and their historical time series data in the historical typhoon dataset; Input the passing path of the typhoon and the time series data of the corresponding passing path in the ocean meteorological model into the preset movement prediction model to obtain a movement action path; Compare the corrected action path and the movement action path to determine the mutation point of the typhoon, and calculate the turning direction and its turning probability of the typhoon at this mutation point based on the ocean meteorological model.

8. A typhoon path integrated prediction system based on ocean meteorological big data, characterized in that, Include: An alignment and fusion module, which is used to collect multi-source marine meteorological data, align and fuse the multi-source marine meteorological data in terms of time and space resolution to obtain a marine meteorological model; A prediction module, which is used to extract typhoon meteorological features from the marine meteorological model, obtain a historical typhoon data set, and predict the initial action path of the typhoon according to the typhoon meteorological features and the historical typhoon data set; A correction module, which is used to detect and obtain the ocean temperature and salinity values on the initial action path, determine the ocean temperature and salinity anomaly values, and correct the initial action path according to the ocean temperature and salinity anomaly values to obtain a corrected action path; A mutation module, which is used to calculate the mutation points and their turning probabilities of the typhoon based on a preset movement prediction model, in combination with the passing path of the typhoon and the marine meteorological model; A deviation correction module, which is used to correct the corrected action path according to the mutation points and their turning probabilities to obtain the final typhoon path, and display the final typhoon path on a geographic information platform.

9. A terminal device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the typhoon path integrated prediction method based on marine meteorological big data as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program, wherein when the computer program runs, it controls the device where the computer-readable storage medium is located to execute the typhoon path integrated prediction method based on marine meteorological big data as described in any one of claims 1 to 7.

Images