ANC Seminar - John Quinn

Tuesday, 5th March 2024

Mapping the World's Buildings with Satellite Imagery

Abstract: 

Building footprints are needed for many important applications, such population estimation, emergency response, urban planning and environmental science. In this talk I'll describe a machine learning approach to identify buildings in high resolution satellite imagery, culminating in the Open Buildings dataset, with 1.8 billion building polygons across Africa, South/Southeast Asia and Latin America. I'll also show some recent results using lower resolution (but frequent and freely available) Sentinel 2 imagery, which demonstrate that by using temporal stacks of images, a much higher level of detail about buildings and roads can be obtained with this source than was previously thought possible, at a global scale.

Bio:

John Quinn is a Staff Software Engineer at Google Research and a Director of Sunbird AI, a research nonprofit based in Uganda. He was previously technical lead for Africa projects at United Nations Global Pulse, and Senior Lecturer in Computer Science at Makerere University in Uganda. He has worked on a number of large scale AI projects across the African continent, in the fields of remote sensing, speech and language, agriculture and health. He holds a BA in computer science from the University of Cambridge (2000), and a PhD in machine learning from the University of Edinburgh (2007).

Event type: Seminar

Date: Tuesday, 5th March

Time: 11:00

Location: G.03

Speaker(s): John Quinn

Chair/Host: Arno Onken