About / News / Full News Story

RIT emerges as leader in drone research
Remote Sensing
Faculty/Staff
Graduate

Mar. 22, 2015
Sean Lahman

(Photo: MAX SCHULTE/@maxrocphoto/ / STAFF PHOTOGRAPHER)

On a cold March afternoon, RJ Garma is flying a small quad copter in a parking lot at the Rochester Institute of Technology. The craft, about 2 feet wide, hovers several hundred feet in the air as Garma controls its movements using an application on his tablet. A camera mounted to the bottom of the copter sends overhead images of the campus back to him.

But Garma is not just a student out for a few hours of fun. He's a U.S. Air Force captain, a doctoral candidate in the school's imaging science program, and one of a handful of RIT students and faculty piloting a revolution.

The school has long been known for its expertise in aerial and satellite photography — a science known as remote sensing. So it's no surprise that RIT has now emerged as one of the world's leading centers for research on drones, small unmanned aircraft.

David Messinger, interim director of the school's Center for Imaging Science, says he gets calls almost every week from companies seeking this expertise, and graduating students are in extraordinarily high demand. Messinger says he can't recall one in the last 10 years who walked across the stage without having a job lined up.

"Our students don't bother going to the job fair," he says, "because when employers want our students they come here and talk to them directly."

The Digital Imaging and Remote Sensing Lab is the biggest group within the center, and about three-quarters of the DIRS students are working on masters and doctorate programs. They're highly sought after by government and industry alike.

The Federal Aviation Administration has also turned to RIT. With drone technology advancing faster than regulators can keep up, the FAA has designated six organizations across the country to conduct research to help devise rules for the operation of unmanned aircraft systems (UAS) in the United States. RIT and the Massachusetts Institute of Technology are the lead institutions for one of those six, a coalition of universities and industry called the Northeast UAS Airspace Integration Research Alliance, or NUAIR.

The development of inexpensive flying platforms — as well as ever smaller and cheaper digital cameras — has suddenly made aerial photography something anyone can do. For a few hundred dollars, you can walk into the mall or go online at Amazon and get a simple drone capable of taking pictures. With a modicum of skill, just about anyone can take overhead pictures of an urban landscape or a natural wonder.

Those breathtaking photos are nice, but what's really driving interest, what's turning this into a science and big business, is the idea of converting those aerial images into useful data.

"If I launched a drone and over the course of half an hour it covered the entire RIT campus at a 1-inch resolution, I'm not going to be able to physically look at all of that data." Messinger explained. "You've got to have some back end processing schemes that try to extract information out of the data.

"That's what you really want. You don't want the pictures. Nobody cares about the pictures. You want the information that you can get out of it," he said.

One of the first areas to leverage this new technology is precision agriculture, and researchers at RIT have been developing systems to address issues like drought management and disease detection.

Here's how it works. A farmer launches a small drone by simply throwing it into the air. The aircraft circles a couple of times to orient itself, then it goes back and forth until it has taken high-resolution pictures of each individual plant in his entire field. When it lands, the information is downloaded to a computer that can begin analyzing the images, asking questions about what the images show.

Questions like: Is that plant healthy? Is there a gap in my irrigation system? Is there a broken pipe someplace or an infestation of something that's moving across the field?

A test program launched in Genesee County last year, testing sensors that can estimate crop yields, spot potential pests or diseases, and help farmers to apply fertilizer more precisely. Another project by an RIT student examined vineyards in the Finger Lakes, using spectral imaging to assess water levels in plants.

Messinger says that the process usually starts with researchers going to customers and saying "this is what we can do, tell us if it's useful." If you could look at each plant in a thousand-acre field every day, what could you learn from that?

Sometimes it is about building these complex applications, but other times it's about developing solutions for simpler problems. Garma has been working on a system that takes off and follows him as he walks around. It flies in circles around him taking pictures. It's essentially a selfie drone, and it's not hard to imagine all sorts of practical applications for this technology.

Carl Salvaggio, an RIT professor who oversees the undergraduate program at the Center for Imaging Science, is working on building functionality into unmanned systems that makes them easier for nonexperts to use. You're not always going to have Ph.D.-level engineers like him to operate these systems, after all.

Salvaggio developed an inexpensive imaging system that can transmit live images of an area about an acre in size. It's designed for law enforcement or first responders who want to get an overhead view of what's going on in real time.

"A user can simply point at a spot on Google Maps," Salvaggio said. "The system will figure out where it is, turn and keep the camera trained toward that point on the ground the user selected."

In addition to the technology push, there's also an application pull: folks who come to RIT with a specific problem they need help solving. And perhaps the biggest problem is the one faced by the Federal Aviation Administration, which is charged with developing a plan for getting drones integrated into the national airspace.

They're concerned about these inexpensive fliers getting in the way of commercial aircraft, of course, and there are all sorts of technical and logistical issues that need to be addressed.

As one of the lead test centers for NUAIR, researchers at RIT are working on solutions for these challenges.

"With manned aircraft, we're pretty good at navigating from point to point," said Agamemnon Crassidis, a professor in RIT's Kate Gleason College of Engineering and the academic director for NUAIR.

Commercial aircraft use sophisticated navigation systems with an array of high-tech sensors.

"Those systems are large and they're expensive. You're not going to put an $80,000 inertial navigation system on a small unmanned aircraft," Crassidis said. "We're trying to develop sensors that are just as accurate but much cheaper, weigh less, use less power, and obviously are a lot smaller."

Collision avoidance is a major concern because drones are flying at lower altitudes than traditional manned aircraft. Crassidis says it's pretty easy to avoid buildings or hillsides, but that smaller obstacles — electrical wires or tree branches — present a more complex challenge. Part of the solution is developing better "detect and avoid" algorithms, but the real advances will be driven by those new sensors.

"The variety of potential applications for these unmanned systems is amazing, but we have to be able to do the testing to figure out how we can do those things safely," Crassidis said. "In terms of the technology, we're pretty close."

SLAHMAN@Gannett.com

Read More Read Full Story »
Original Source: Democrat and Chronicle