A team of students led by researchers in the Energy Access Project and the Energy Data Analytics Lab will apply machine learning techniques to high resolution aerial imagery data to identify the location of solar panels throughout Cape Town, South Africa. Currently, solar panels are being used by wealthier households as backup electricity sources during power outages. This work will result in a reference dataset of key features that has not yet been defined in this context as well as a model that will be able to quantify the locations and sizes of solar panels in different neighborhoods. This initiative will further research at Duke that is dedicated to understanding the role of clean backup energy in South Africa.
Project Lead: Marc Jeuland
Project Manager: Biz Yoder