Artificial Intelligence Model for Predicting Cyclone Intensity

Abstract

By utilizing transfer leaming methodologies and INSAT 3D satellite imagery, the objective of this study is to develop a system for detecting cyclone intensity. Cyclones, one of the most destructive natural phenomena, cause severe damage to both lives and property. Efficient disaster management and mitigation efforts are contingent upon the timely identification and accurate assessment of cyclone intensity_x000D_ The geostationary weather satellite INSAT 3D generates high-definition images depicting the surface of the earth. These photographs effectively depict the atmospheric conditions that are conducive to cyclone detection, Clouds, noise, and other atmospheric disturbances make it challenging to distinguish cyclones with precision from these images._x000D_ In this study, we propose a cutting-edge transfer learning method for cyclone Intensity detection. Transfer learning is a machine learning technique that permits the modification of previously trained models to be applicable to various problem domains. In order to extract relevant features from the INSAT 3D satellite images, we employ convolutional neural networks (CNNs) that have undergone training on large datasets since their inception. These attributes are utilized by the classifier to identify and categorize the cyclones according to their intensity._x000D_ We conducted experiments using a substantial dataset of INSAT 3D satellite images in order to evaluate the performance of the proposed method. The results indicate that the proposed methodology achieves a 95% overall accuracy in cyclone detection. Furthermore, the methodology exhibits a notable level of precision when it comes to approximating cyclone strength.