Map features are point data that have been extracted from Mapillary imagery using computer vision technology. You can view these features on the Mapillary web app or download them to use for your own mapping and databasing needs.
This article will cover everything you need to know about map features!
How are map features generated?
Map features are directly related to object detections—if an object has been detected in multiple Mapillary images, we are able to reconstruct its location in three-dimensional space and assign it positional coordinates. By assigning each object a set of coordinates (latitude and longitude values), we are able to place it on a map.
To learn more about the computer vision principle behind our map features, triangulation, see the Wikipedia article on Triangulation.
What types of map features can Mapillary detect?
Mapillary map features include traffic signs and other common objects associated with vehicular and pedestrian experiences—think fire hydrants, benches, lane markings, street lights, utility poles, et cetera.
We split map features into two main categories: points and traffic signs. Mapillary currently supports 42 point features and 1,500 classes of traffic signs. Due to the sheer amount of traffic sign types, and a need to differentiate between them, we like to think of traffic signs as their own category when referring to map features.
How can I view map features?
You can explore map features on the Mapillary web app via the 'Map data' pop-up. To learn more about how to view and filter map features in the web app, see our articles on exploring point features and traffic signs, respectively.
How can I download map features?
To integrate Mapillary map features into your own maps, databases, and workflows, you'll need to download them or call them with our API.
Downloading map features as a GeoJSON is possible via the Mapillary web app, the Mapillary API, or the Mapillary Python SDK.
Downloading with the Mapillary web app is best for a small focus area—think a town or a city neighborhood. If you need to download map features for a larger area, we recommend using the Mapillary API.
An example of a Python script to download map features can be found here.
Tips for high-quality map features
The quality of Mapillary map features is tied directly to the quality of the imagery used to generate them. If your reason for uploading imagery to Mapillary is primarily for extracting map features, please be mindful of the following:
- To generate a map feature, an object must be detected in at least two (for traffic signs) or three (for all other point features) images. We cannot attempt to position an object without multiple data points on it. Keep this in mind when choosing the interval of the images you capture—more images mean more data points to triangulate the position of the object more accurately.
- The accuracy of a map feature position is influenced by the accuracy of image locations. This is contingent on the GPS accuracy of your phone, camera, or other GPS device. If high-accuracy positioning is a requirement, you might want to consider professional devices with enterprise-grade GPS capabilities.