About the event
Understanding the spatially-embedded energy system is necessary to manage generation intermittency, to mitigate climate risks and associated social impacts, and to target optimal policy interventions. Remote sensing and computer vision offers a novel method for localising energy infrastructure agnostic to company and country reporting. Using a longitudinal corpus of remote sensing imagery and machine learning we provide a globally-exhaustive inventory of utility-scale solar PV generating stations, complete with installation dates for facilities built after June 2016. We also present ongoing research in the development of a self-supervised sensor-fusion model for general-purpose semantic embedding of remote sensing imagery.