Crowdsourcing of map images
| Reference number | |
| Coordinator | Mapillary AB |
| Funding from Vinnova | SEK 1 720 000 |
| Project duration | December 2013 - August 2017 |
| Status | Completed |
Purpose and goal
Today, Mapillary has collected 188M images in 190 countries and extracted over 14 billion map objects using computer vision algorithms. Many global mapping providers use the service for improving maps and 2017 saw the beginning of real customer revenue.
Expected results and effects
One goal of the project was to see if it was possible to generate quality map data from images collected by anyone, using e.g. smartphones. This has been validated and we can report that such data is now available and automatically generated by the company at scale world-wide. The effects of this is that we help enable smarter cities, safer vehicles, and better maps.
Planned approach and implementation
The crowdsourced model picked up momentum and quickly generated lots of data and validation. Technically generating map data from these images took longer time than expected, but succeeded in the end. B2B licensing of images and data, was the outcome of the business model exploration.