Detecting Tropical Cyclone Using Infrared Satellite Images

Abstract

Tropical Cyclone (TC) are among the most destructive natural disasters. Analysis of TC image from infrared satellite images is an active field of research. Many algorithms were developed in past few decades on TC image analysis. The location of TC is always an important and difficult problem. Many researchers have tried to detect the right Region of Interest (ROI) from the image automatically. In this paper, feature extraction method based on modified Morphological processing and the color segmentation approach based on the intensity transformation and color spaces are applied for automatically extraction the location of storms. 45 TCs occurred between 1989 and 2014, are tested and several experiments have been done to evaluate the proposed system.