Deep Learning Regression and Classification of Tropical Cyclones based on HURSAT data

Abstract

This study focuses on the application of Convolutional Neural Network (CNN) models for cyclone intensity estimation and classification. Cyclones are one of the most devastating natural disasters that cause significant loss of life and property damage. Accurate and timely estimation and classification of cyclones can help in effective disaster management and mitigation. In this study, a CNN Lenet architecture is used for classification and a proposed CNN model is used for intensity estimation of cyclone satellite images. The method involves pre-processing the data, training the models on the dataset, and evaluating the model's performance. The results demonstrate that the models achieve high accuracy in cyclone classification and low Root Mean Square Error (RMSE) in cyclone intensity estimation. This study provides insights into the potential of CNN architecture, which can be beneficial in enhancing disaster management strategies. © 2024 URSI.