Thunderstorm strong wind grade prediction classification method based on multi-source convolutional neural network

Abstract

The invention discloses a thunderstorm strong wind grade prediction classification method based on a multi-source convolutional neural network. According to the method, a multi-source convolutional neural network model is adopted to carry out feature extraction on various data images obtained by the Doppler meteorological mine, more meteorological data information can be fused, and the extractionof difference features is improved. Meanwhile, the method is combined with a classification method in a support vector machine, and a model obtained on a meteorological data training set of small andmedium samples has a very good thunderstorm strong wind level prediction classification effect.

Description

technical field [0001] The invention relates to the fields of image processing and disaster weather prediction, in particular to a local thunderstorm and gale level prediction method based on a multi-source convolutional neural network and a support vector machine. Background technique [0002] my country is one of the countries with frequent occurrence of strong convective weather, heavy precipitation, hail, thunderstorms, strong winds and tornadoes. With the development of the economy, the losses caused by the occurrence of strong convective weather will be more serious. Severe convective weather is one of the difficulties and key points in the current weather forecasting business due to its small time-space scale, rapid change, severe weather, large social impact, and complicated occurrence and development mechanism. [0003] The monitoring of severe convective weather is an important part of severe convective weather forecasting, especially the basis of short-term nowca...

AI Technical Summary

Problems solved by technology

The cross-entropy loss function will make the model overfit. In many cases, the intra-class distance of the extracted feature vector is greater than the inter-class distance.

The existing methods also lack the processing of samples that are not easy to classify by the softmax classifier, and for small sample data sets, the prediction of the softmax classifier is prone to overfitting, and it is difficult to improve the robustness of the prediction classification

Images