Generating the Cloud Motion Wind Field from Satellite Cloud Imagery Using Deep Learning Approach

Abstract

Cloud motion wind (CMW) are routinely derived by tracking features in sequential geostationary satellite infrared cloud imagery. In this paper, we explore the cloud motion wind algorithm based on the data-driven deep learning approach. And different from conventional hand-craft feature tracking and correlation matching algorithms, we use deep learning model to automatically learn the motion feature representations and directly output the field of cloud motion wind. In addition, we propose a novel large-scale cloud motion wind dataset (CMWD) for training deep learning models. We also try to use a single satellite cloud imagery to predict the cloud motion wind field in a fixed region, which is impossible to achieve using traditional algorithms. The experimental results demonstrate that our algorithm can predict the cloud motion wind efficiently, even with a single cloud imagery as input.