We noticed you're browsing in private or incognito mode.

To continue reading this article, please exit incognito mode or log in.

Not a subscriber? Subscribe now for unlimited access to online articles.

  • Mapillary
  • Connectivity

    Crowdsourced maps should help driverless cars navigate our cities more safely

    Swedish startup Mapillary is compiling a huge database of roadside objects such as signs and markings to help driverless vehicles get around.

    Our current street maps aren’t much good for helping driverless cars get around. Although we’ve mapped most roads, they get updated only every couple of years. And these maps don’t log any roadside infrastructure such as road signs, driveways, and lane markings. Without this extra layer of information, it will be much harder to get autonomous cars to navigate our cities safely. Robotic deliveries, too, will eventually require precise details of road surfaces, sidewalks, and obstacles. 

    A Swedish startup called Mapillary thinks it has the answer. It’s an open platform that crowdsources images of streets taken by people on their smartphones: a sort of Wikipedia of mapping. It says it is now one of the biggest publicly available databases of street-level imagery in the world.

    Sign up for The Download
    Your daily dose of what's up in emerging technology

    “Driverless cars need the latest view of the road,” says CEO Jan Erik Solem. “They require a higher and higher update frequency for maps, from quarterly to monthly to weekly to daily. The only scalable way to do that is using technology.”

    There are a number of approaches to gathering mapping data, and there’s competition among startups within the field. Mapillary says it is different because it is crowdsourced—unlike, say, Google’s proprietary Street View, which is updated every couple of years. Because anyone can contribute to its platform, it gets updated every day.

    Image of Mapillary street-view map

    This approach is similar to OpenStreetMap, which launched in 2004 and provides a free, editable map of the world, but it doesn’t log any of these additional roadside details.

    Mapillary uses computer-vision software to analyze the crowdsourced images and identify objects. Its database of 422 million images covers 6.2 million kilometers (3.9 million miles) of the globe. And it’s growing all the time: its software has just put 186 million objects, like utility poles, benches, and manhole covers, on the map, locked to a specific location with coordinates.

    Once the images are uploaded, the software analyzes them for objects and identifies them. The maps are then made available for anyone to see online. The service is free for charities and for educational or personal use, but commercial customers have to pay a license fee.

    Mapillary is already being put to use. The city of Amsterdam, transport officials in Vermont and Arizona, and even the country of Lithuania are using the firm’s maps to improve their understanding of their streets.

    “If you look at any object on a street, someone is responsible for it. It all has to be catalogued and checked,” says Steven Hewett, who works for the city of Clovis in New Mexico.

    Photo of driver using Mapillary

    Clovis is using Mapillary to make sure it’s fulfilling its obligations to residents by keeping driveways clear, signs up to date, and roads clear of potholes. The city used to do this by paying a contractor to go around and painstakingly log every mailbox, fire hydrant, and stop sign, for a few dollars each.

    It has now automated virtually the entire process using Mapillary. “Without this software, we’d be walking around collecting all the data by hand, and I can’t even guess how long that’d take for a 23-square-mile city,” Hewett says.

    Gabriel Brostow, a computer science professor at University College London, agrees that this kind of mapping needs to be automated to be scalable. “Millions of square kilometers in the world can’t be updated by humans as quickly as they could be by algorithms,” he says.

    One day, driverless cars could be both consumers and producers of this data, capturing street-level imagery as they travel (both faces and license plates are automatically blurred out by Mapillary’s software). Indeed, Hewett foresees a future where networked vehicles can automatically identify issues like fallen trees in the road, or traffic accidents, and automatically notify the relevant public authorities.

    The data could also help equip cities with better knowledge of their streets so they can improve public transport and accessibility for disabled people, Brostow adds.


    *This story was corrected to make clear the images are crowdsourced but the underlying code is not open source.

    Learn from the humans leading the way in connectivity at EmTech Next. Register Today!
    June 11-12, 2019
    Cambridge, MA

    Register now
    Image of Mapillary street-view map
    Photo of driver using Mapillary
    More from Connectivity

    What it means to be constantly connected with each other and vast sources of information.

    Want more award-winning journalism? Subscribe to Print + All Access Digital.
    • Print + All Access Digital {! insider.prices.print_digital !}*

      {! insider.display.menuOptionsLabel !}

      The best of MIT Technology Review in print and online, plus unlimited access to our online archive, an ad-free web experience, discounts to MIT Technology Review events, and The Download delivered to your email in-box each weekday.

      See details+

      12-month subscription

      Unlimited access to all our daily online news and feature stories

      6 bi-monthly issues of print + digital magazine

      10% discount to MIT Technology Review events

      Access to entire PDF magazine archive dating back to 1899

      Ad-free website experience

      The Download: newsletter delivery each weekday to your inbox

      The MIT Technology Review App

    You've read of three free articles this month. for unlimited online access. You've read of three free articles this month. for unlimited online access. This is your last free article this month. for unlimited online access. You've read all your free articles this month. for unlimited online access. You've read of three free articles this month. for more, or for unlimited online access. for two more free articles, or for unlimited online access.