Today we announced that Mapillary has joined Facebook with our missions to map the world converging. This is a significant step forward in the Mapillary journey which many of you have been an instrumental part of. Our decision to join Facebook allows us to build upon our efforts to map the world. We’ll have the resources and scale to help support the collection of street-level imagery while applying computer vision that will greatly increase the quantity and quality of map data available for OpenStreetMap.
The OpenStreetMap community has been an integral part of Mapillary since we began, so I’d like to share our perspective on what this change means for OpenStreetMap and why we’re so excited for the next step in Mapillary’s journey.
A collaborative model
Since Mapillary began in 2013, we have sought to build a collaborative platform where images and computer vision are used to map the world. The OpenStreetMap community was quick to embrace Mapillary and see the potential street-level images have to build better maps. Here are just a few examples of the community has been using street-level imagery:
- Ballerup, Denmark: neogeografen has uploaded millions of images in Ballerup and surrounds which he uses to map cycle paths, street-lighting, and points of interest.
- Dar es Salaam, Tanzania: The Humanitarian OpenStreetMap Team and the Ramania Huria project made use of Trimble’s advanced 360º camera to collect imagery and map flood resilience characteristics.
- Xayaboury, Laos: The World Bank has been working with Laos’ Department of Transportation to build and monitor roads in Xayaboury province. Imagery helps ensure the roads are mapped in OpenStreetMap during and after the construction phase.
- Padova, Italy: The state of urban accessibility was mapped using 10,000 images and meetings with citizens and associations. This project proved the viability of data collection, analysis, and policy advocacy with OpenStreetMap and Mapillary as primary tools.
Mapillary platforms and licensing
As we outlined in the blog post, our plan is to continue being a global platform for imagery, map data, and improving all maps. You’ll be able to continue capturing, uploading, and working with Mapillary images.
All of the images and data on Mapillary are available for editing OpenStreetMap. Nothing changes here and Mapillary imagery and data will continue to be available for OpenStreetMap in compliance with the OpenStreetMap Foundation Contributor terms.
Mapillary + Facebook for OSM
Mapillary’s involvement in OpenStreetMap has been one of the highlights of our journey to date. Here is a quick summary of the tools we’ve built and/or supported during this time.
- Street-level imagery integration in iD Editor
- Mapillary plugin for JOSM
- Traffic sign layer added to iD Editor and JOSM
- Map feature layer added to map features
- Advanced image filtering in JOSM
Mapillary’s image and traffic sign layers shown within JOSM
This next chapter of Mapillary’s journey is an opportunity to build upon our OpenStreetMap efforts to a degree that was not possible earlier on. Previous OpenStreetMap development had to be balanced against development work for other mapping use cases. As part of Facebook, we can focus on OpenStreetMap while supporting Facebook’s open mapping efforts.
OpenStreetMap helps to power location-centric experiences at Facebook, such as their Marketplace that drives transactions for millions of small businesses in addition to supplying vital data to humanitarian organizations around the world. Facebook improves OpenStreetMap through map editing and Map with AI, an open AI-powered set of services/tools that includes the RapiD editor, a fork of iD Editor. RapiD demonstrates the potential of AI-extracted map features when they’re made available to the community in a workflow that increases mapping speed, and more importantly, drive a higher level of detail, quality and accuracy on the map. As part of Facebook, we are combining our experiences applying computer vision to street-level and satellite imagery. This enables us to build exciting tools for OpenStreetMap, empowering the community to capture and use imagery in new ways.
Mapillary’s AI derived map features viewed within iD Editor
We’re incredibly grateful for the support and vision the OpenStreetMap community has shown since the early days of Mapillary. As we join the Facebook team, we hope you will continue to shape our thinking, informing and utilizing tools that together allow us to create an open map of the world.