Identifying Urban Makeshift Communities using satellite imagery and geo-coded data
summary
Describes using transfer learning and deep computer vision on recent satellite imagery combined with geo-coded data to automatically detect non-permanent urban poor (makeshift) communities in Bangalore. Covers an applied Machine Learning pipeline developed by DataKind Bangalore to replace Pollinate Energy’s manual Google Maps scouring and door-to-door discovery, with the goal of reducing field work and resource use.
AI-written summary · about
chapters (4)
related talks
48mUsing Artificial Intelligence and Satellite Imagery to Zero In on the Philippines’ Most Vulnerable Communities
58mImproving disaster response with machine learning
07mMapping Urban Deprivation
06mInterview with David Soloff
46mDeep Learning for Satellite Image Processing, Lessons Learned
33mWhat and where? - Machine learning for geospatial image analysis
10mMapping sidewalks and street furniture with AI and computer vision
18mDeep learning for Space imagery applications