MCL251: An Eye in the Sky: How Radiant Solutions Processes Satellite Imagery with AI and Amazon Mechanical Turk

This planet is a big place filled with amazing and unusual things. Understanding every object, location, and action on this pale blue dot is an enormous challenge. As the world's leading provider of high-resolution Earth imagery, data, and analysis, DigitalGlobe faces this challenge every day. They use automatic computer vision and machine learning where possible, but so far, the only true solution requires the most powerful information processing machine we know: the human brain. Scaling this solution to work on the trillions of satellite pixels collected by DigitalGlobe every day requires thousands of brains, all working in harmony. To address this, DigitalGlobe | Radiant's (now Radiant Solutions) Tomnod service uses Amazon Mechanical Turk, a crowdsourcing internet platform, to identify small objects appearing in large areas of new satellite imagery. Tomnod is heavily used for commercial and humanitarian purposes. In this session, you hear how Radiant Solutions uses crowdsourcing to help solve large-scale computer vision and machine learning problems.