Storm-Net: Advancing Cyclone Intensity Forecasting with Deep Learning on INSAT 3D IR Imagery

Accurate estimation of tropical cyclone intensity has gained increasing importance, particularly in recent years, for effective disaster preparedness and response. We explore important developments in this area, emphasizing the use of satellite imagery and deep learning methods. TCICENet (Zhang et al., 2021), LSTM-CSO (Kumar et al., 2023), and TCNN (Chen and Yu, 2021) are three examples of recent studies that show how deep learning can be used to improve cyclone intensity accuracy. These studies emphasize the integration of additional data sources, optimization techniques, and a variety of datasets. Challenges linked to abrupt intensity fluctuations and low-intensity cyclones are identified, pointing to areas for further advancement. Exploring future research directions, including adapting deep learning models across geographical regions and their pivotal role in bolstering early warning systems, is a key aspect of our analysis. This research is vital for mitigating the impact of cyclones on vulnerable communities and we are in the process of creating a deep learning model to estimate the intensity of tropical cyclones, drawing upon CNNs, LSTMs, and optimization techniques, thus enhancing early warning systems and community resilience.