Product Overview: Mapillary
Mapillary is a comprehensive platform designed to make street-level images and map data accessible and usable on a global scale, facilitating the automation and enhancement of mapping processes.
What Mapillary Does
Mapillary brings together a global network of contributors who capture and upload street-level images using a variety of devices, including smartphones, 360° cameras, action cameras, dashcams, and professional rigs. These images are then processed using advanced computer vision and machine learning technologies to extract valuable map data. The platform aims to make the world more accessible by visualizing it through immersive street-level views and providing detailed map data, which can be used for various applications such as urban planning, navigation, and autonomous driving.
Key Features and Functionality
Image Contribution and Processing
- Anyone can contribute to the platform by uploading images from any type of camera. This collaborative approach has resulted in over two billion images contributed from 190 countries since its inception in 2013.
Automatic Data Extraction
- Mapillary’s sophisticated algorithms automatically extract map data from the uploaded images, including the identification and positioning of up to 42 types of objects such as utility poles, streetlights, manholes, and trash cans. Initially focused on traffic sign recognition, the platform now detects over 1,500 types of traffic signs and additional object classes.
Imagery Viewer and Comparison
- The platform offers an imagery viewer that combines images into views that can be explored and compared across different time periods. This feature allows users to see changes and updates in real-time, enhancing the accuracy of maps.
Privacy Protection
- To ensure privacy, Mapillary uses technology to automatically blur faces and license plates detected in the images. Manual editing tools are also available for additional privacy adjustments.
Cloud Access and Sharing
- All street-level imagery and map data are accessible and shareable via the cloud, enabling users to access and utilize the data from anywhere, free of charge.
Object Detection and Segmentation
- The platform automatically identifies and segments imagery into various categories of assets, providing a detailed breakdown for in-depth analysis. This includes full segmentation of imagery and the detection of specific objects, allowing for filtering images based on the presence of these objects.
API and Integration
- Mapillary provides an API that allows users to query and access the extracted map data, including object detections and map features. This facilitates integration with other tools and applications, making it a powerful resource for geospatial analysis and urban planning.
Collaboration and Open Data
Mapillary is built on the principle of collaboration and open data. By making the data available to everyone, the platform encourages a community-driven approach to mapping, which helps in keeping maps updated, improving city infrastructure, enhancing mobility and navigation, and supporting autonomous driving initiatives.
In summary, Mapillary is a robust platform that leverages user-contributed street-level images, advanced computer vision, and machine learning to create and update detailed maps. Its key features include automatic data extraction, privacy protection, cloud access, object detection, and a collaborative open-data approach, making it an invaluable tool for various mapping and urban development applications.