Mapillary works on the principle that anyone can capture and use street-level images. By sharing data on a collaborative platform, everyone wins. As the need for up-to-date maps keeps growing, computer vision is the key to scaling up map data generation. Here is a short summary of how the system works.
Anyone can create an account and contribute images to Mapillary. All you need for that is your smartphone and our mobile app. You can also use other cameras, as long as the images get geotagged. You capture sequences of images and upload them to Mapillary.
After uploading, your images will be processed and connected with everyone else’s images. Faces and license plates are blurred for privacy. Our computer vision technology creates 3D reconstructions and recognizes objects in the images. Mapillary shares several bits of the processing pipeline as open source, for example, OpenSfM and MapillaryJS. (When something is open source, it means that the code with which a system is built is available for anyone to modify, use and/or distribute.)
The images on Mapillary are available for anyone to explore. The images and the data extracted from them can be used by map makers, cities, transport agencies, logistics and delivery companies, automotive companies, humanitarian organizations, NGOs, and anyone else who needs geospatial information from street-level images.