Artificial Intelligence for Satellite Image Processing: Application to Rainfall Estimation

Meteorological satellites are widely used for collecting information about the atmosphere. Due to the indirect relation between satellite data and measurements of meteorological parameters, mathematical models, especially those based on artificial intelligence, have been developed for meteorological modeling. Indeed, in recent years, machine learning has enabled fundamental advances in the modeling of random systems. In this context, we will show the contribution of techniques based on artificial intelligence in the estimation of precipitation. Based on the expertise of our research laboratories in this field, the objective of this paper is to present our recent results and developments using machine learning, such as ANN, SVM, and RF. For the classification and estimation of rainfall intensities, satellite images were used for the implementation of these techniques. The training and validation was carried out by comparing the satellite images to the corresponding radar images. The results of these artificial intelligence-based techniques indicate very interesting performance.