An intelligent optimized cyclone intensity prediction framework using satellite images

Weather prediction is the hottest topic in remote sensing to understand natural disasters and their intensity in an early stage. But in many cases, the typical imaging models have resulted in less forecasting rate. Hence, to overcome this problem, a novel buffalo-based Generalized Adversarial Cyclone Intensity Prediction System (BGACIPS) was designed for cyclone intensity prediction using space satellite images. The processed satellite images contained features like rain, snow, Tropical depression (T.Depression), thunderstorms (T.strom), and cyclone. Initially, the noise features were removed in the pre-processing module, and the refined data was entered into the classification layer. Consequently, the analysis of the features was performed, and the intensity of each feature and cyclone stages were identified. Furthermore, the planned design is executed in the python environment, and the improvement score has been analyzed regarding prediction exactness, mean errors, and error rate. Hence, the proposed novel BGACIPS has a lower error rate and higher prediction accuracy than the compared models.