Aircraft

ScanEagle UAV and image recognition software used to track seals

View 2 Images
Scientists are using a Scan Eagle UAV and image recognition software to track seals on the diminishing Arctic ice cap (Photo: Boeing)
Seals on an ice flow, as picked out by the Boulder Labs image recognition system (Photo: CIRES)
Scientists are using a Scan Eagle UAV and image recognition software to track seals on the diminishing Arctic ice cap (Photo: Boeing)

The Arctic region is currently experiencing a warming trend which is seeing the ice cap break apart, and may even ultimately result in the total absence of ice in the summer months. Many scientists attribute this trend to man made global warming, but whatever the cause, it’s not good news for the seals that breed, rest, and escape marine predators on the ice. In an effort to understand the full scope of the situation, scientists have turned to the Boeing-designed Scan Eagle – an unmanned aerial vehicle more often used for military reconnaissance.

The research project is being led by Elizabeth Weatherhead of the Cooperative Institute for Research in Environmental Sciences (CIRES), a joint venture of the University of Colorado at Boulder and the National Oceanic and Atmospheric Administration (NOAA). Her team is using UC Boulder’s Scan Eagle to conduct aerial photographic surveys of ice floes in the Bering Sea. Launched from the NOAA research vessel McArthur II, the camera-equipped 10-foot (3-meter) wingspan UAV has been making 3 to 5 mile (5 to 8 km) runs lasting 2 to 8 hours over the sea ice, at altitudes of 300 to 1,000 feet (91.5 to 305 meters).

The resulting photographs are analyzed to see how many seals are present, and on what types of ice – “By finding the types of ice they prefer, we can keep track of that ice and see how it holds up as the Arctic sea ice extent shrinks,” said Weatherhead.

Seals on an ice flow, as picked out by the Boulder Labs image recognition system (Photo: CIRES)

Given that it can be very Where’s Waldo-ish trying to find the seals within over 27,000 photos, Boulder Labs has created image recognition software that automatically picks out seals in the pictures. The software also helps researchers in identifying the species of seals, while future versions could additionally deduce their age and gender, and could be tweaked to look for polar bears and their tracks.

“Biologists are thrilled about the image recognition software because it could change the way we monitor seal populations,” said Weatherhead. “We can send an unmanned craft out from a ship, collect 4,000 images, and have them analyzed before dinner. This is a great example of physicists working closely with biologists who are concerned with the health of seal populations.”

  • Facebook
  • Twitter
  • Flipboard
  • LinkedIn
2 comments
Facebook User
these UAVs are great . . you can\'t tell me that our government can\'t catch illegals as they cross our borders or spot illegal traffic off our shores . . . we don\'t have to spend billions of dollars for this kind of protection . . .
Facebook User
Borders really aren\'t lines on a map anymore. Los Angeles is the 2nd largest Mexican City after Mexico City. It is about 60 miles north of the line on the map we want to call the border.
Until the Gold Rush of 1849, California was first part of Mexico and then an independent republic. Only when Gold Rush news reached the east coast did California become a State. Motivation is pretty obvious, no?.
As to drones, I was in LA area the other night and saw a couple accidents on I-210. What should have been a two hour drive around LA took six hours. Helicopters and ambulances everywhere. Maybe 2-3 people were killed? Who cares? LA area has between 16-18 million people depending on how you define \"illegal.\" The \"border\" goes far into the central valley counting the workers at Fast Food places near I-5. Either we white guys face up to reality of Mexico/California or Mexicans will do it for us.