How Mapillary supports OpenStreetMap Communities

�December 2019

Mapillary is the street-level imagery platform that scales and automates mapping

Scaling and automating maps with cameras

Problem

Traditional maps struggle to keep up with demands of new mobility��Solution

• Scale by opening up data collection to any camera, anywhere
• Automate with computer vision to speed up map data extraction�

Status

Converting imagery to map data at scale is now a necessary core component in mapping and automotive.

Data collection for organizations

Any camera, anywhere

Mapillary Organization

Create a new organization

Switch to your new organization

Adding team members to your organization

Add people to your organization

Assign roles

Create a shape for Capture Project

Biella Capture Project

Capture Project toolset

Organize image collection with your team

• Follow progress in real time as the drivers use a simple app to complete their routes—capturing with smartphones or other cameras.

Dense capture example

Processing the images to generate map data

Processing the imagery

800 million images

99.9% of faces and license plates anonymized

Verification:

Human in the loop

Mapillary Verification project

False Positives

Verification tool new version

Map data extraction

Geolocated traffic signs

1,500 traffic sign classes from 100 countries

Geolocated street assets

42 object types globally

Object detections

65 object classes globally

Filter imagery by object type and export image locations to view on a map

Also available with computer vision generated annotations

​

​

​

​

Map editing

Making map data available in OpenStreetMap

​

ID Editor

​

ID Editor

​

JOSM Plugin

​

Pic4Review

​

Map data request

Thank you!

@mapillary

said@mapillary.com

​