One of the most potent and devastating types of weather phenomena on Earth are tropical
cyclones (TC). A cyclone is a large air mass that revolves in a powerful epicenter of low
atmospheric pressure, rotating counterclockwise in northern hemisphere and clockwise in
southern hemisphere. In this work, a unique deep learning model for predicting the path of
tropical cyclones is proposed. Unlike previous predictions, cyclones have the potential to cause
serious harm to both individuals and their possessions. To lessen the adverse impacts of
cyclones, better and more precise prediction systems are required. In order to anticipate
cyclones, this work proposes an efficient Rain Optimized Convolutional Neural Network (RO
CNN). The input image is manipulated and preprocessed using Notch filter. Grey Level
Coordination Matrix (GLCM) is utilized for gathering the features from segmented output after
the processed image and it is fed into K-means clustering technique for segmentation. When
compared to other conventional methods, the proposed classifier achieves higher levels of
accuracy in classification. In order to validate its effectiveness, suggested ROA-CNN based
system is implemented in MATLAB platform and accuracy obtained during classification
using ROA-CNN is about 95.8%.