Massachusetts Institute of Technology (MIT) researchers have developed software that can identify houses in rural villages from satellite images, potentially saving time that would otherwise be spent sending teams from village to village.
The researchers' initial projects focused on villages in India and Africa. The first project selected villages in sub-Saharan Africa for a program of unrestricted cash grants to help people in low-income rural areas improve their standard of living by enabling them to buy equipment, livestock, and other goods. The system selected the poorest villages by counting the percentage of houses with thatched roofs compared with those topped by more-expensive metal roofs. The second project selected villages in rural India for the installation of microgrids to supply electricity from solar panels and battery-storage systems, and then determined the best sites for those panels and the most efficient network layout for distributing the power.
For both projects, the process begins by having users visually examine the satellite images and pick out the houses. Those examples are then entered in as training data for a machine-learning system, which generalizes the criteria for determining what is a house and what is not. Over time, the program can learn from the hand-picked data and get better at determining where the houses are, according to MIT researcher Brian Spatocco.
From MIT News
View Full Article
Abstracts Copyright © 2015 Information Inc., Bethesda, Maryland, USA