In this project, we propose to leverage the growing number of highly detailed multi-spectral images of the Sun to improve the prediction of the solar wind streams arriving at the Earth. We propose to exploit the capabilities of modern computer vision and machine learning techniques to register solar images, analyze them and assimilate them in an empirical (data-driven) model. We plan to develop and train a regression model that predicts solar wind based on the large amount of available data (solar wind measurements near the Earth and multi-spectral solar images): the amount of data transmitted by SDO alone is approximately 1TB per day. Huge amounts of historical observations are now available from NASA and associated centers, whereas a large database of solar images from various missions is maintained by GFZ-Potsdam.