Powered by Indico
v3.2.9
Soil maps are used across a wide range of disciplines to determine optimal or sustainable land-use strategies, assess environmental impacts, and understand climate change. Nonetheless, at a time when climate change and sustainability are at the forefront of global politics, across much of the world and particularly in the global south, the quality of soil maps is relatively poor. An interdisciplinary team from Durham's Geography and Archaeology departments aims to harness AI and remote sensing techniques to develop, train and test a prototype deep convolutional network capable of mapping soil from Earth observation data. If successful, our method could be used to improve soil mapping around the world, including in regions where modern conflicts prevent soil mapping techniques that rely on fieldwork.