Dr Alireza Taravat

Dr Alireza Taravat

Dr Alireza Taravat
Dr Alireza Taravat

Dr Alireza Taravat, a remote sensing scientist at Kiel University, Germany, explains how he used Meteosat data as part of a project.

Last Updated

05 March 2021

Published on

04 March 2016

Clouds are one of the most important meteorological phenomena affecting the Earth radiation balance and satellites are extremely useful tools for helping to understand such phenomena.

Accurate information about the physical and radiative properties of clouds is essential in order to determine the role of clouds in the climate system and for retrievals of surface and aerosol properties.

Spotlight on Users — Dr Alireza Taravat
Band combination of SEVIRI images over Europe and multiple-layer perceptron neural networks classifier result

I started to work on EUMETSAT products when I was doing my PhD project on the application of machine learning techniques in earth observation, at Tor Vergata University.

In the last year of my PhD, the aim of my work was to explore how machine-learning techniques can explore the Earth's atmosphere and surface, with an emphasis on cloud detection.

I decided to examine the machine learning algorithms capabilities for automatic cloud detection in whole-sky images. After getting good results and publishing them (See Neural Networks and Support Vector Machine Algorithms for Automatic Cloud Classification of Whole-Sky Ground-Based Images, registration needed), it was the time to move to the next step and create a machine learning model for automatic classification of clouds on MSG SEVIRI data.

MSG SEVIRI data are captured from a geostationary orbit, so they can produce more images of the Earth in a given time than is possible from satellites in other orbits, and the spectral capabilities (12 spectral bands) of MSG allows an accurate cloud cover analysis.

Because of those reasons I have decided to choose MSG SEVIRI data for my research — Multilayer Perceptron Neural Networks Model for Meteosat Second Generation SEVIRI Daytime Cloud Masking. This image is an example of the results (the first column of the figure shows the band combination of SEVIRI images over Europe and multiple-layer perceptron neural networks classifier result has been shown in the second column).

However, we are always searching for improvements to the data that we use, and looking to the future with the Meteosat Third Generation (MTG) series of satellites that will allow us to gain an even better understanding of the Earth and atmosphere.