The U.S. National Geospatial Intelligence Agency is calling on the private sector to develop machine-learning tools to automate repetitive and time-consuming image analysis tasks.
For example, researchers from the Center for Geospatial Intelligence at the University of Missouri have used a deep-learning neural network to assist human analysts in visual searches for surface-to-air missile sites on a large area in southeastern China. The study found the system achieved an average search time of 42 minutes for an area of about 90,000 square kilometers, a result the researchers say is more than 80 times more efficient than a traditional human visual search.
In addition, the team notes the software achieved the same overall statistical accuracy human analysts--90%--for correctly locating missile sites.
Meanwhile, the researchers note artificial intelligence can be used for data mining to help prioritize information so networks are not clogged by data that could be valuable.
From Space News
View Full Article
Abstracts Copyright © 2017 Information Inc., Bethesda, Maryland, USA
No entries found