Research On Typhoon Track Prediction Based On Spatial-Temporal Sequence Features

Typhoons,as a typical tropical weather system,are accompanied with strong winds and heavy rains,which also trigger secondary disasters such as floods,landslides,mudslides and storm surges,which pose a great threat to the protection of natural environment and people’s life and property safety.The outcome of typhoon warning information can timely help the government to take emergency disaster prevention measures,remind residents adjust travel strategies,avoid danger in time,and minimize casualties and economic losses.However,there are many factors that affect typhoon tracks,and traditional typhoon track prediction lacks the power to mine nonlinear features in high-dimensional data,failing to deeply explore the impact of influencing factors on changes in typhoon tracks.In response to the above issues,the paper proposes a typhoon track prediction model based on spatio-temporal sequence features,which fully utilizes multidimensional historical monitoring data,excavates the temporal and spatial features of various dimensions of typhoon track data,and achieves high-precision prediction of various forms of typhoon tracks.The main work content is as follows:(1)Sea surface meteorological observation and statistical product dataset fusion and analysis.Collect the best track data of tropical cyclones and air-sea data,fuse heterogeneous data from different sources,then clean the data and perform normalization processing,analyze and screen the principal components of air-sea data,and finally generate a fusion dataset containing spatio-temporal characteristics.(2)A typhoon tracking prediction model based on spatial-temporal sequence features is proposed.Firstly,convolutional neural network was used to extract the spatial features of the input data,and then long and short term memory neural network was used to extract the temporal features of the input data.Finally,the attention mechanism was used to fuse the spatial features and temporal features,and the relationship between the sea and air data and the dimensions of the optimal track data set of tropical cyclones was deeply excavated.(3)Improving the gray wolf optimization algorithm.The grey Wolf optimization algorithm is introduced and its convergence factor is improved.Combined with the typhoon track prediction model proposed in this paper,the parameter seeking process of the model is optimized to improve the training efficiency and prediction accuracy of the model.(4)Designing and implementing the typhoon warning system.Combined with the actual needs of typhoon track prediction,the functions of the typhoon early warning system are analyzed and designed,including typhoon early warning,typhoon track display,data download and other functions,and the application and interface visualization of the typhoon track prediction model are realized.The typhoon track prediction model proposed in this paper makes full use of the temporal and spatial characteristics of the typhoon track and realizes the real-time prediction of the typhoon track.Meanwhile,grey Wolf optimization algorithm is introduced to improve the parameter optimization process of the model,effectively improving the training efficiency of the model and the prediction accuracy of the typhoon track.It provides a feasible idea and method for typhoon track prediction and typhoon early warning,which has high theoretical research significance and practical application value.