The Map Layer Your Sensors Cannot Infer
Your vehicles already see the road. What they cannot see is what the road means: which lane connects to which, where the crossing is, which sign governs the junction ahead. MapAtlas produces that layer as data, for every area you operate in.
Built From Imagery, Not Inherited From a Database
Satellite, ground and street-level imagery, reconciled into one picture
Public map data is uneven. A junction mapped in detail in one city is a bare line in the next, and coverage thins exactly where operators are expanding. We combine satellite imagery, ground and street-level imagery and existing map data to produce the layer for a given area, rather than passing on whatever happens to already exist there.
One Schema, Every City You Operate In
The same street is modelled differently depending on where it is. European data tends to record a pavement as an attribute of the road; North American data draws it as separate geometry. Sign conventions diverge the same way. Operators running several regions end up building and maintaining a separate extractor for each one. We normalise it, so the data behaves identically in every market.
Delivered in the Formats Your Stack Already Reads
Bulk extracts per operating region by default, exported to Lanelet2, ASAM OpenDrive, GeoJSON or PBF. Where a runtime dependency suits you better, the same data is reachable through our routing, map matching and isochrone APIs.
What the Layer Contains
Road network and lane connectivity. Sidewalks, footways and crossings. Kerbs and drop-kerbs. Traffic signs, signals, stop and give-way control. Speed limits. Points of interest and address data where your application needs them.
Frequently Asked Questions
Tell Us Where You Are Driving
Every deployment has its own operating design domain, its own formats and its own expansion plan. The useful first conversation is about where you run today and where you are going next.