Segmentation of tropical cyclone eye using satellite infrared images

Abstract

The tropical cyclones are destructive weather systems and are known for their devastating effects during landfall. Cyclone tracking is one of the important tasks for the meteorologist. The eye of the tropical cyclone is the most remarkable feature. The eye of the cyclone is the roughly circular area extending over 30-65 km in diameter. The deepest convection is found around the eyewall for some tens of kilometers. The eye grows deeper when the cyclone becomes heavy and the winds speed grows high. In this study, the data from the 1995-2016 of the CIRA imagery for the tropical cyclone of the Bay of Bengal basin is analyzed and the model is developed to determine the eye of the cyclone. The segmented eye features are fed into the Rule Based Classifier which classifies the tropical cyclone images based on the presence and absence of the eye.