Artificial intelligence systems for rainy areas detection and convective cells' delineation for the south shore of Mediterranean Sea during day and nighttime using MSG satellite images

Abstract

The aim of this study is to investigate the potential of cloud classification by means of support vector machines using high resolution images from northern Algeria. The images were taken from the Spinning Enhanced Visible and Infrared Imager (SEVIRI) on board of the Meteosat Second Generation (MSG) satellite. An automatic system was developed to operate during both day and nighttime by following two steps of data processing. The first aims to detect rainy areas in cloud systems, whereas the second delineates convective cells from stratiform ones. A set of 12 spectral parameters was selected to extract information about cloud properties, which are different from day to night. The training and validation steps of this study were performed by in-situ rainfall measurement data, collected during the rainy season of years 2011 and 2012 via automatic rain gauge stations distributed in northern Algeria. Artificial neural networks (ANNs) and support vector machine (SVM) were explored, by combining spectral parameters derived from MSG images. Better performances were obtained by the SVM classifier, in terms of Critical Success Index and Probability of Detection for rainy areas detection (CSI=0.81, POD=91%), and also for convective/stratiform delineation (CSI=0.55, POD=74%).