Towards a Dynamic Production of Global Land Cover Maps

<!--HTML--><p>Land cover (LC) maps are a fundamental tool in a variety of fields such as climatology, ecology and geography, as their availability is crucial for the analysis of trends and recognition of patterns of phenomena that occur on the Earth’s surface. The dense Time Series of images with a worldwide coverage provided by current satellite missions allow the dynamic production of LC maps, and several methods have been recently proposed to generate maps at country, continental or global scale. Despite recent advances made possible by the deployment of Deep Learning methods on satellite data, there are still a number of open challenges to generate accurate and consistent LC maps. The aim of this talk is to introduce the problems that are lying in front of researchers in this multidisciplinary field, offering an overview on the available approaches and pointing to possible solution toward frequent updates of LC maps on a global scale.<br><i><span style="mso-ansi-language:EN-GB;mso-fareast-language:EN-GB;">Rocco comes from Venice, Italy and received his Master's Degree in Information Engineering at the University of Trento in 2019. His interests lie mainly in Deep Learning applied to Remote Sensing data and Distributed Deep Learning.</span></i><br><strong>Coffee will be served at 10:00.</strong></p>