InstaRoad
Project Showcase

InstaRoad

Extracting Roads from Medium Resolution Satellite Imagery

By: Kelvin Wei , Catherine Li , Jing Yeh

Supervised by: Patrick Marais , Jonathan Shock


About

Abstract

We develop AI tools to monitor how road networks change over time, based on satellite imagery. This supports urban planning and assessing the environmental impact of road networks. While satellite imagery makes this possible at scale, high-resolution imagery is expensive, and licensing often restricts model training. To overcome this, we use freely available Sentinel imagery, which covers the globe every few days. However, this data is low-resolution, and this needs to be accounted for in the algorithms. 

Documents 2

Downloadable resources and documentation

Click "View Full" to open documents in a new window

Gallery 1

Explore the visual story of this exhibit