How much better is a computer at finding surface-to-air missiles compared to trained, human analysts? Eighty times faster, according to a study by University of Missouri researchers. The Center for Geospatial Intelligence (CGI) used a deep-learning neural network to spot air-defense sites over a wide area of southeastern China using imagery from DigitalGlobe. The network completed the task in 42 min., whereas it took people 60 hr. Both methods were 90% accurate. The report, published in the ...

THIS CONTENT REQUIRES SUBSCRIPTION ACCESS

You must have an Aviation Week Intelligence Network (AWIN) account or subscribe to this Market Briefing to access "Speeding The Search For Surface-to-Air Missiles".

 

Current Aviation Week Intelligence Network (AWIN) enterprise and individual members: please go to http://awin.aviationweek.com for access.

 

Not currently a subscriber? Click on the "Learn More" button below to view subscription offers.

Already registered? here.