Skip to main content

MIT’s new A.I. could help map the roads Google hasn’t gotten to yet

RoadTracer: Better Automated Maps

Google Maps is a triumph of artificial intelligence in action, with the ability to guide us from one place to another using some impressive machine learning technology. But while the routing part of Google Maps doesn’t need too many humans in the mix, manually tracing the roads on the aerial images to make them machine usable is incredibly time-consuming and mundane. As a result, even with thousands of hours spent on this task, Google employees still haven’t managed to map the majority of the 20 million-plus miles of roadways that stretch around the world.

Fortunately, researchers from the Massachusetts Institute of Technology’s Computer Science and Artificial Intelligence Laboratory (CSAIL) and Qatar Computing Research Institute may have come up with a solution. They developed an automated method to build roadmaps which is 45 percent more accurate than existing methods. Called RoadTracer, the work uses neural networks to intelligently map roads on images. The system could be especially well-suited to map parts of the world where maps are frequently out of date, like remote and rural areas in the developing world.

“We trained the neural net using aerial images of 25 cities across six countries in North America and Europe,” Favyen Bastani, a graduate student at MIT CSAIL, told Digital Trends. “Specifically, for each city we assembled a corpus of high-resolution satellite imagery from Google Earth and ground truth road network graphs from OpenStreetMap, covering a region of roughly 10 square miles around the city center.”

RoadTracer works by starting with a known location on a road network and then examining the surrounding area to work out what is most likely to be the next part of the road. Once this point has been added, the process is repeated again and again until the entire road network has been added.

Going forward, the team hopes to move beyond relying on principally aerial images for mapping. “For example, they don’t give you information about roads with overpasses, since you obviously cannot see them from above,” Bastani said. “One of our other projects is to train systems on GPS data, and then to eventually be able to merge these approaches into a single mapping system.”

A paper describing the work will be presented in June at the Conference on Computer Vision and Pattern Recognition (CVPR) in Salt Lake City.

Editors' Recommendations

Luke Dormehl
I'm a UK-based tech writer covering Cool Tech at Digital Trends. I've also written for Fast Company, Wired, the Guardian…
Futuristic new appliance uses A.I. to sort and prep your recycling
Lasso

Lasso: The power to change recycling for good

Given the potential planet-ruining stakes involved, you’d expect that everyone on Earth would be brilliant at recycling. But folks are lazy and, no matter how much we might see footage of plastic-clogged oceans on TV, the idea of sorting out the plastic, glass, and paper for the weekly recycling day clearly strikes many as just a little bit too much effort.

Read more
MIT’s clever robotic basketball hoop will help you level up your game
MIT robot basketball hoop

Researchers from Massachusetts Institute of Technology’s Computer Science & Artificial Intelligence Laboratory might not seem like the folks most likely to help you improve your hoop-shooting basketball skills. But that’s exactly what a new MIT CSAIL project sets out to do with a quirky basketball-training machine featuring a basketball hoop that shrinks and raises when you make shots, thereby shape-shifting to help improve the various facets of your game.

For example, to begin with, the basket can be positioned at a lower height with wider hoop diameter, which gradually shrinks down and also raises to reach regulation proportions as you score more and more baskets. It’s an unusual project from a lab that’s more used to working with the latest cutting-edge artificial intelligence algorithms -- but, as it turns out, it fits perfectly with CSAIL’s areas of expertise.

Read more
This clever new technique could help us map the ocean floor — from the sky
pass ocean mapping stanford airbornesonarconcept

A friend of mine who works in games design recently showed me a 3D model of the Earth, rendered in great detail using topographically accurate satellite data, so that we could soar through canyons and our respective neighborhoods at high speed like a pair of joyriding Supermen. “Let’s see if we can go underwater,” he said, exhilarated, as we flew out over the Pacific.

We couldn’t. The model, so stunningly accurate on land, apparently had zero data with which to model the undersea environment. It was an unrendered void beneath the water’s glassy surface, as if this was some subaquatic version of The Truman Show, and we had reached the end of the world.

Read more