FOSS4G 2023

Mapillary is an open platform for street-level imagery and map data that began in 2013. Since then around 1.8 billion images have been contributed from around the world. Imagery has been contributed from horseback in Kyrgyzstan, boats in the canals of Amsterdam, and bicycles on the streets and trails of Sydney. As Mapillary approaches 2 billion images, we’d like to summarize the latest features, acknowledge some of the amazing contributions, and hint at some of the updates that are coming.

Some of the things that we have been working on include:

Desktop Uploader improvements including support for videos and popular cameras.
Improvements to Mapillary Tools, command line scripts for working with and uploading geotagged imagery and video.
Mobile app updates including multi-tasking, redesigns, multi-language support, and upload improvements.
Camera Grant programs in the US and Europe, providing 360º cameras for people interested to map pedestrian infrastructure.
Integrations with Rapid Editor, an AI powered OpenStreetMap editor which we will demo in more detail at a workshop.
Updated Help Pages to make capturing, uploading, and using street-level imagery far easier.

After walking through the latest Mapillary improvements, we will take a look at case studies of organizations contributing and using imagery. We’ll zoom in on an NGO, a government agency, and a commercial entity, each of which are using Mapillary in different ways.

We’ll finish our talk with an exploration of upcoming Mapillary features and projects. We encourage questions and suggestions in the Q&A and hope for a productive conversation at the end as we walk together towards 2 billion images.

Use cases should drive product development, not the other way around. Maps and the products we use to consume them have the biggest impact on the world when these principles are adhered to. How many government portals have you visited where a carefully curated map is presented that hardly anyone sees let alone uses? Presenting the data to the user in an intuitive way that helps them make a decision or take action is essential.

Large paper maps of the 1700s were well suited to a captain’s desk as their ships traversed the oceans. Road atlases of the 20th century helped to spur family adventures and weekend getaways as highway networks were constructed around the world. The small computers in our pockets today allow us to see when the next train will arrive and which one will get us home sooner. These examples took the technology of the day and used it to make products with significant impact on society. The mobile internet in particular changed mapping in one of the most notable ways since humans started abstracting 3D space on 2D surfaces.

We’re on the cusp of another great shift in the way maps are used with many exciting use cases awaiting discovery. The technology powering this potential is Augmented Reality (AR). This talk will explore some of the use cases that AR is supporting and where it might be useful in future. We’ll look at how AR can be accessed and how the medium of access affects its utility. With these use cases in mind, we’ll assess how open tools and map data enable AR. Some of the data and tools we’ll look at include:
Geometries of pedestrian ways
Associated attributes: Incline, safety, lighting, access, surface type, accessibility features
Building entrances
3D building data
VPS for localisation
Routing algorithms

The talk will conclude with a summary of Meta’s approach to map building and how open source geospatial technology powers the maps we build for today and the years ahead.