Urban Grammar- Form, Function, Culture and Satellite Imagery
The MapScaping Podcast - GIS, Geospatial, Remote Sensing, earth observation and digital geography
00:00
How to Train a Supervised Learning Algorithm
CNN's John Defterios asked a team of researchers to create an algorithm that could look at satellite data. They used what they called "signatures" which were created by dividing the Earth into 16 geographic areas and labeling them based on form and function. The model was able to identify sprawl, for example, using this signature class as well as other labels such as 'sparkle' or 'dismal patches,' he said.
Transcript
Play full episode
