Typhoon prediction method and device, computer equipment and storage medium
By using time-series satellite images and historical trajectory information, combined with target network model and attention mechanism, the problem of insufficient accuracy of typhoon intensity and trajectory prediction is solved, and accurate prediction of typhoon trajectory and intensity is achieved, and the accuracy of prediction results is improved. .
Patent Information
- Application Number
- CN202411340900.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-24
- Publication Date
- 2025-01-07
AI Technical Summary
The prior art lacks accuracy in typhoon intensity and trajectory prediction, which cannot meet the current meteorological service needs for accuracy and real-time.
By obtaining the typhoon's timing satellite image and historical trajectory information, the target network model, including encoder and decoder, combines the temporal attention mechanism and self-attention mechanism, predict the current and future intensity and trajectory of the typhoon.
Accurate prediction of typhoon trajectory and intensity is achieved, the relationship between typhoon trajectory and intensity is fully explored, and the accuracy of prediction results is improved.
Smart Images
Figure CN119270388A_ABST
Abstract
Description
Technical Field
[0001] The present application belongs to the technical field of meteorological forecasting, and in particular, relates to a method, apparatus, computer equipment and storage medium for typhoon forecasting. Background Art
[0002] According to the "China Meteorological Disaster Yearbook", typhoons are one of the most serious natural disasters and are prone to cause tsunamis, floods and other secondary disasters, resulting in heavy casualties and huge economic losses. Determining typhoon path, intensity and other parameters is the first step in making typhoon forecasts and issuing typhoon warnings, which has important application value and research significance. Typhoons are weather systems with complex dynamic mechanisms. Typhoon numerical forecasts require a high level of prior knowledge from forecasters. At present, there is a widespread lack of accuracy in typhoon intensity estimation and trajectory prediction, which makes it impossible to meet the current demand for accurate and real-time meteorological services. Summary of the invention
[0003] The embodiments of the present application provide a method, apparatus, computer device and storage medium for typhoon prediction, aiming to solve the technical problem that the existing technology is not accurate enough in predicting the intensity and trajectory of typhoons.
[0004] In a first aspect, an embodiment of the present application provides a method for typhoon prediction, the method comprising:
[0005] Obtain the time-series satellite images and historical trajectory information of the typhoon at time t;
[0006] The typhoon intensity of the typhoon at time t, the future typhoon intensity at time t+n, and the future typhoon trajectory are predicted based on the time-series satellite images and the historical trajectory information.
[0007] In one embodiment, the predicting the typhoon intensity of the typhoon at time t, the future typhoon intensity at time t+n, and the future typhoon trajectory based on the time-series satellite images and the historical trajectory information includes:
[0008] estimating the typhoon intensity of the typhoon at time t based on the time-series satellite images and the historical trajectory information;
[0009] The future typhoon intensity and future typhoon trajectory of the typhoon at time t+n are predicted based on the time-series satellite images, the historical trajectory information and the typhoon intensity at time t.
[0010] In one embodiment, the method further comprises:
[0011] Inputting the time-series satellite images of the typhoon and the historical trajectory information into a target network model, wherein the target network model includes an encoder and a decoder;
[0012] The estimating the typhoon intensity of the typhoon at time t based on the time-series satellite image and the historical trajectory information includes:
[0013] Encoding the time-series satellite image by the encoder to obtain a first image feature;
[0014] Connecting the first image feature with the historical trajectory information to obtain a first connection feature;
[0015] Decoding the first connection feature through the LSTM layer of the decoder to obtain a fusion feature;
[0016] The fused features are processed through the fully connected layer FC of the target network model to estimate the typhoon intensity of the typhoon at the tth moment.
[0017] In one embodiment, the processing of the fused features through the fully connected layer of the target network model to estimate the typhoon intensity of the typhoon sequence at the t-th moment further includes:
[0018] The typhoon intensity at the t-th moment is stored in a typhoon historical intensity prediction feature set, and the typhoon historical intensity prediction feature set is configured to store the typhoon intensity at each moment predicted by the target network model.
[0019] In one embodiment, the future typhoon intensity of the typhoon at time t+n is predicted by the following method, including:
[0020] Each typhoon intensity in the historical intensity prediction feature set is input into the target network model to obtain future typhoon intensity prediction features to predict the future typhoon intensity of the typhoon at the t+nth time.
[0021] In one embodiment, the future typhoon track of the typhoon at time t+n is predicted by the following method, including:
[0022] Obtaining trajectory coordinate features based on the historical trajectory information through a temporal attention mechanism;
[0023] Connecting the typhoon intensity at the t-th moment with the future typhoon intensity prediction feature to obtain a second connection feature;
[0024] The future typhoon track of the typhoon at time t+n is predicted based on the second connection feature and the track coordinate feature.
[0025] In a second aspect, an embodiment of the present application provides a typhoon prediction device, the device comprising:
[0026] An acquisition module is used to obtain the time-series satellite images and historical trajectory information of the typhoon at time t;
[0027] A prediction module is used to predict the typhoon intensity of the typhoon at time t, the future typhoon intensity at time t+n, and the future typhoon trajectory based on the time-series satellite images and the historical trajectory information.
[0028] In one embodiment, the prediction module further includes a current intensity estimation module, a future intensity estimation module and a trajectory estimation module;
[0029] The current intensity estimation module is used to estimate the typhoon intensity of the typhoon at time t based on the time-series satellite image and the historical trajectory information;
[0030] The future intensity estimation module is used to input the intensity of each typhoon into the target network model to obtain the future typhoon intensity prediction characteristics to predict the future typhoon intensity of the typhoon at the t+nth time;
[0031] The trajectory estimation module is used to obtain trajectory coordinate features based on the historical trajectory information through a temporal attention mechanism; connect the typhoon intensity at the tth moment and the future typhoon intensity prediction feature to obtain a second connection feature; and predict the future typhoon trajectory of the typhoon at the t+nth moment based on the second connection feature and the trajectory coordinate feature.
[0032] In a third aspect, the present invention further proposes a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor; when the processor executes the computer program, the steps of the typhoon prediction method described in the first aspect above are implemented.
[0033] In a fourth aspect, the present invention further proposes a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor; when the processor executes the computer program, the steps of the typhoon prediction method described in the first aspect above are implemented.
[0034] The beneficial effect of the present application is that, based on the time-series satellite images and historical track information of the typhoon, the track and intensity of the typhoon can be accurately predicted at the same time, the relationship between the track and intensity of the typhoon can be fully explored, and the accuracy of the prediction results can be improved. BRIEF DESCRIPTION OF THE DRAWINGS
[0035] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0036] Figure 1 is a structural diagram of a computer device provided in one embodiment of the present application;
[0037] Figure 2 This is a schematic flow chart of the main steps of a typhoon prediction method provided by the present application;
[0038] Figure 3 It is a structural diagram of a framework for typhoon track estimation involved in an embodiment of the present application;
[0039] Figure 4 This is a schematic flow chart of a specific embodiment of a typhoon prediction method provided in this application;
[0040] Figure 5 This is a structural block diagram of the typhoon testing device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0041] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.
[0042] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.
[0043] It should also be understood that the term “and / or” used in the specification and appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.
[0044] As used in the specification and appended claims of this application, the term "if" can be interpreted as "when" or "uponce" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "uponce it is determined" or "in response to determining" or "uponce [described condition or event] is detected" or "in response to detecting [described condition or event]", depending on the context.
[0045] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.
[0046] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.
[0047] An embodiment of the present invention provides a computer device, such as Figure 1 As shown, Figure 1 The computer device 1 of this embodiment includes: at least one processor 1001 ( Figure 1 Only one is shown), a memory 1005, and a computer program 1006 stored in the memory 1005 and executable on the at least one processor 1001. When the processor 1001 executes the computer program 1006, the steps in the method embodiment of typhoon prediction of the present application are implemented.
[0048] Those skilled in the art will understand that Figure 1 This is merely an example of the computer device 1 and does not constitute a limitation on the computer device 1 , which may include more or fewer components than shown in the figure, or a combination of certain components, or different components.
[0049] The processor 1001 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.
[0050] In addition, the computer device 1 also includes a satellite image acquisition interface 1004, which can, for example, access a professional meteorological website, or start a professional meteorological software or application program, to obtain time-series satellite images of the typhoon;
[0051] The computer device 1 also includes a track information acquisition interface 1003 for acquiring historical track information of typhoons. For example, professional meteorological websites may be accessed, or professional meteorological software or applications may be started to acquire track information of typhoons at various times.
[0052] In some embodiments, the memory 1005 may be an internal storage unit of the computer device 1, such as a hard disk or memory of the computer device 1. In other embodiments, the memory 1005 may also be an external storage device of the computer device 1, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the computer device 1. Further, the memory 1005 may also include both an internal storage unit of the computer device 1 and an external storage device. The memory 1005 is used to store an operating system, a computer execution program of the typhoon prediction method, etc.
[0053] It is understandable that due to the complexity of the atmosphere and the spatiotemporal characteristics of meteorological data, typhoon track and intensity prediction is very challenging. How to effectively use the spatiotemporal information contained in the sequence is a key issue, and the track and intensity prediction of the typhoon have a certain intrinsic connection, but the existing technical research methods do not combine the two. The typhoon prediction method and prediction device proposed in the embodiment of the present application can fully explore the relationship between the typhoon track and intensity, and can effectively estimate and predict the intensity and track of the typhoon.
[0054] refer to Figure 2In order to solve the above technical problems, the main steps of the typhoon prediction method provided in the embodiment of the present application are steps S10-S20:
[0055] Step S10, obtaining time-series satellite images and historical track information of the typhoon at time t;
[0056] In the specific implementation, the computer device can access professional meteorological websites, or start professional meteorological software or applications to obtain the time-series satellite images and historical trajectory information of the typhoon at the tth moment, where t can be understood as the current moment or any moment. Satellite remote sensing data covers a wide area and has high temporal resolution. Satellite cloud images contain important information on the formation, development, and extinction of typhoons, and are currently the main means of typhoon observation and forecasting.
[0057] Step S20, predicting the typhoon intensity of the typhoon at time t, the future typhoon intensity at time t+n, and the future typhoon trajectory based on the time-series satellite images and the historical trajectory information, wherein n represents the length of time from the tth time to the t+nth time.
[0058] Specifically, the present application will use the target network model to predict the typhoon intensity at time t, the future typhoon intensity at time t+n, and the future typhoon trajectory. The time-series satellite image of the typhoon and the historical trajectory information are input into the target network model, wherein the target network model includes an encoder and a decoder.
[0059] It is understandable that the target network model of the embodiment of the present application is a network model of an encoder-decoder fused with an attention mechanism. Based on the target network model, the embodiment of the present application can simultaneously predict the typhoon track and intensity, and then fully explore the relationship between the typhoon track and intensity.
[0060] In the specific implementation, in order to further improve the accuracy of the prediction, this embodiment uses temporal attention and self-attention mechanisms to mine spatiotemporal relationships. The attention mechanism can effectively improve the target network model's ability to discriminate features.
[0061] The framework of typhoon track estimation is as follows Figure 3 As shown, the framework of this embodiment uses an encoder-decoder structure based on a recurrent neural network (RNN) and predicts the current typhoon intensity, future intensity, and future trajectory based on typhoon satellite images and the longitude and latitude information of the typhoon center. For example, given a typhoon i, a picture sequence observation of length m+1 and the corresponding trajectory coordinates Among them, l t Represents the longitude and latitude coordinates {x, y} of the typhoon center at time t.
[0062] The embodiment of the present application can accurately predict the track and intensity of a typhoon based on the time-series satellite images and historical track information of the typhoon, fully explore the relationship between the track and intensity of the typhoon, and improve the accuracy of the prediction results.
[0063] In order to further Figure 2 The typhoon prediction method is described in detail. Figure 4 As shown, the above step S20 in the embodiment of the present application further includes step S21 and step S22:
[0064] Step S21: estimating the typhoon intensity of the typhoon at time t based on the time-series satellite images and the historical trajectory information, including sub-steps S211 to S214:
[0065] Sub-step S211, encoding the time series satellite image by the encoder to obtain a first image feature F1;
[0066] Sub-step S212, connecting the first image feature F1 with the trajectory coordinate feature of the corresponding time in the historical trajectory information to obtain a first connection feature;
[0067] Sub-step S213, decoding the first connection feature through the LSTM layer of the decoder to obtain a fusion feature;
[0068] Sub-step S214, processing the fused features through the fully connected layer FC of the target network model to estimate the typhoon intensity int of the typhoon at the tth moment.
[0069] It is understandable that the cyclic neural network structure based on the encoder and decoder proposed in the embodiment of the present application estimates the typhoon intensity. The encoder includes a convolutional layer to extract image features and a ConvLSTM layer to extract the relationship between the image sequence. The input is a picture sequence. The first image feature F1 encoded by the encoder is connected to the longitude and latitude geographic information of the typhoon center (historical trajectory information) to obtain a more discriminative feature representation (first connection feature). The decoder includes an LSTM layer. The decoded features (fused features) are used to estimate the (current) typhoon intensity int at the tth moment through the fully connected layer FC. Figure 3 , Figure 3 The framework can estimate the wind circle radius, the maximum wind speed in the center of the typhoon and other intensity parameters, and use the cross entropy loss function for training.
[0070] It should be noted that after the fusion features are processed by the fully connected layer of the target network model to estimate the typhoon intensity of the typhoon sequence at the t-th time, the method further includes: storing the typhoon intensity at the t-th time in the typhoon historical intensity prediction feature set {s t ,st-m+1 ,...,s t-m}, the typhoon historical intensity prediction feature set {s t ,s t-m+1 ,...,s t-m} is configured to store the typhoon intensity s at each time (t, t-m+1,…tm) predicted by the target network model.
[0071] Step S22: predicting the future typhoon intensity and future typhoon trajectory of the typhoon at time t+n according to the time-series satellite image, the historical trajectory information and the typhoon intensity at time t, wherein the future typhoon intensity of the typhoon at time t+n is predicted by sub-step S22A, and the future typhoon trajectory of the typhoon at time t+n is predicted by sub-steps S22B1, S22B2, and S22B3:
[0072] Sub-steps S22B1 to S22B3 (predicting future typhoon tracks):
[0073] Sub-step S22B1: obtaining trajectory coordinate features based on the historical trajectory information through a temporal attention mechanism;
[0074] Sub-step S22B2: The typhoon intensity int at the tth moment and the future typhoon intensity prediction feature S are combined through the fully connected layer. pred ={s t+1 ,s t+2 ,...,s t+n} are connected to obtain a second connection feature;
[0075] Sub-step S22B3: predicting the future typhoon trajectory of the typhoon at time t+n based on the second connection feature and the trajectory coordinate feature.
[0076] It is understandable that the trajectory prediction objective function of the embodiment of the present application is the predicted trajectory distribution p(L pred |L obs ,S pred ,Int i ).in is the estimate of the center position of typhoon i at time t+1, is the observation of the center position of typhoon i. pred ={s t+1 ,s t+2 ,...,s t+n} is the prediction characteristic of future typhoon intensity. tis the current typhoon intensity. The model input is the observation of the typhoon position at time t and the corresponding image sequence, and the output is the prediction at time t+n. When predicting the trajectory, the embodiment of the present application uses two attention mechanisms. The encoder input uses the time attention mechanism and the decoder input uses the self-attention mechanism. The time attention mechanism can select the key frames with the most effective information from the observation sequence. The self-attention mechanism acts on the features learned by the encoder to focus on the feature fragments that are most effective for the current prediction. Finally, the trajectory coordinates at the corresponding moment are predicted through the fully connected layer. The MSE loss is used to train the network during the training stage. The MSE loss is defined as:
[0077]
[0078] In the specific implementation, the trajectory coordinate features are obtained through the time attention mechanism and LSTM, and the FC layer is used to integrate the current intensity prediction feature int and the future typhoon intensity prediction feature S pred The LSTM and FC layers are then used to predict future typhoon trajectories.
[0079] Sub-step S22A (predicting future typhoon intensity): the historical intensity prediction feature set {s t ,s t-m+1 ,...,s t-m The typhoon intensities in the target network model are input into the target network model to obtain the future typhoon intensity prediction feature S pred ={s t +1 ,s t+2 ,...,s t+n}, to predict the future typhoon intensity of the typhoon at time t+n.
[0080] It is understandable that the framework of intensity prediction is similar to that of trajectory prediction, and the main goal is to predict p(S pred |S obs ). Where S obs is the intensity sequence up to time t. The intensity prediction framework also uses the temporal attention mechanism and the self-attention mechanism to learn more discriminative features, and finally uses the MSE loss function to train the framework.
[0081] In the specific implementation, the typhoon historical intensity prediction feature set {s t ,s t-m+1 ,...,s t-m} is used as input and sent to the bidirectional LSTM module and FC module to obtain the future typhoon intensity prediction feature S pred ={s t+1 ,s t+2 ,...,s t +n}.
[0082] The typhoon prediction method proposed in the embodiment of the present application can simultaneously predict the trajectory and intensity of a typhoon, fully explore the relationship between the typhoon trajectory and intensity, and use the attention mechanism to more effectively learn the spatiotemporal information of the typhoon to estimate and predict the intensity and trajectory of the typhoon.
[0083] refer to Figure 5 , Figure 5 A structural block diagram of a typhoon testing device provided in an embodiment of the present application, comprising an acquisition module 01 and a prediction module 02;
[0084] Acquisition module 01, used to obtain the time-series satellite images and historical track information of the typhoon at time t;
[0085] The prediction module 02 is used to predict the typhoon intensity of the typhoon at time t, the future typhoon intensity at time t+n, and the future typhoon trajectory based on the time-series satellite images and the historical trajectory information.
[0086] The prediction module 02 also includes a current intensity estimation module 021, a future intensity estimation module 022 and a trajectory estimation module 023;
[0087] The current intensity estimation module 021 is used to estimate the typhoon intensity of the typhoon at time t based on the time-series satellite image and the historical trajectory information;
[0088] The future intensity estimation module 022 is used to input the intensity of each typhoon into the target network model to obtain the future typhoon intensity prediction characteristics to predict the future typhoon intensity of the typhoon at the t+nth time;
[0089] The trajectory estimation module 023 is used to obtain trajectory coordinate features based on the historical trajectory information through a time attention mechanism; connect the typhoon intensity at the tth moment and the future typhoon intensity prediction feature to obtain a second connection feature; and predict the future typhoon trajectory of the typhoon at the t+nth moment based on the second connection feature and the trajectory coordinate feature.
[0090] It should be noted that the above device can be understood as Figure 1 The virtual device in the computer device 1, the computer device 1 executes the functions of each module of the typhoon prediction device provided in the embodiment of the present application by running the processor 1001; the information interaction, execution process and other contents between the above-mentioned devices / units are based on the same concept as the method embodiment of the present application, and their specific functions and technical effects brought about can be specifically referred to the method embodiment part, which will not be repeated here.
[0091] In addition, an embodiment of the present application further provides a storage medium, which is a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the above-mentioned method embodiments can be implemented.
[0092] If the integrated unit 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 this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, which can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium can at least include: any entity or device that can carry the computer program code to the camera / terminal device, recording medium, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, RandomAccess Memory), electric carrier signal, telecommunication signal and software distribution medium. For example, a USB flash drive, a mobile hard disk, a magnetic disk or an optical disk. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electric carrier signals and telecommunication signals.
[0093] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.
[0094] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.
[0095] In the embodiments provided in the present application, it should be understood that the disclosed devices / network equipment and methods can be implemented in other ways. For example, the device / network equipment embodiments described above are merely schematic. For example, the division of the modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.
[0096] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0097] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.
Claims
1. A method for typhoon prediction, characterized in that: The method comprises: Obtain the time-series satellite images and historical trajectory information of the typhoon at time t; The typhoon intensity of the typhoon at time t, the future typhoon intensity at time t+n, and the future typhoon trajectory are predicted based on the time-series satellite images and the historical trajectory information, wherein n represents the length of time from the tth time to the t+nth time.
2. The method according to claim 1, characterized in that The predicting of the typhoon intensity at time t, the future typhoon intensity at time t+n, and the future typhoon trajectory of the typhoon based on the time-series satellite images and the historical trajectory information includes: estimating the typhoon intensity of the typhoon at time t based on the time-series satellite images and the historical trajectory information; The future typhoon intensity and future typhoon trajectory of the typhoon at time t+n are predicted based on the time-series satellite images, the historical trajectory information and the typhoon intensity at time t.
3. The method according to claim 2, characterized in that The method further comprises: Inputting the time-series satellite images of the typhoon and the historical trajectory information into a target network model, wherein the target network model includes an encoder and a decoder; The estimating the typhoon intensity of the typhoon at time t based on the time-series satellite image and the historical trajectory information includes: Encoding the time-series satellite image by the encoder to obtain a first image feature; Connecting the first image feature with the trajectory coordinate feature of the corresponding time in the historical trajectory information to obtain a first connection feature; Decoding the first connection feature through the LSTM layer of the decoder to obtain a fusion feature; The fused features are processed through the fully connected layer FC of the target network model to estimate the typhoon intensity of the typhoon at the tth moment.
4. The method according to claim 3, characterized in that After the fusion feature is processed by the fully connected layer of the target network model to estimate the typhoon intensity of the typhoon sequence at the t-th moment, the method further includes: The typhoon intensity at the t-th moment is stored in a typhoon historical intensity prediction feature set, and the typhoon historical intensity prediction feature set is configured to store the typhoon intensity at each moment predicted by the target network model.
5. The method according to claim 4, characterized in that The future typhoon intensity of the typhoon at the time t+n is predicted by the following methods, including: Each typhoon intensity in the historical intensity prediction feature set is input into the target network model to obtain future typhoon intensity prediction features to predict the future typhoon intensity of the typhoon at the t+nth time.
6. The method according to claim 5, characterized in that The future typhoon track of the typhoon at the time t+n is predicted by the following methods, including: Obtaining trajectory coordinate features based on the historical trajectory information through a temporal attention mechanism; Connecting the typhoon intensity int at the tth moment and the future typhoon intensity prediction feature to obtain a second connection feature; The future typhoon track of the typhoon at time t+n is predicted based on the second connection feature and the track coordinate feature.
7. A typhoon prediction device, characterized in that: The device comprises: An acquisition module is used to obtain the time-series satellite images and historical trajectory information of the typhoon at time t; A prediction module is used to predict the typhoon intensity of the typhoon at time t, the future typhoon intensity at time t+n, and the future typhoon trajectory based on the time-series satellite images and the historical trajectory information.
8. The device according to claim 7, characterized in that The prediction module also includes a current intensity estimation module, a future intensity estimation module and a trajectory estimation module; The current intensity estimation module is used to estimate the typhoon intensity of the typhoon at time t based on the time-series satellite image and the historical trajectory information; The future intensity estimation module is used to input the intensity of each typhoon into the target network model to obtain the future typhoon intensity prediction characteristics to predict the future typhoon intensity of the typhoon at the t+nth time; The trajectory estimation module is used to obtain trajectory coordinate features based on the historical trajectory information through a temporal attention mechanism; The typhoon intensity at the t-th moment is connected with the future typhoon intensity prediction feature to obtain a second connection feature; and the future typhoon trajectory of the typhoon at the t+n moment is predicted based on the second connection feature and the trajectory coordinate feature.
9. A computer device, characterized in that: The computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor; when the processor executes the computer program, the steps of the typhoon prediction method described in any one of claims 1 to 6 are implemented.
10. A storage medium, wherein the storage medium is a computer-readable storage medium storing a computer program, and when the processor executes the computer program, the method steps of typhoon prediction according to any one of claims 1 to 6 are implemented.