mapbox-vision-ar-android
The Android Vision AR SDK from Mapbox. It uses on-device neural networks to turn a connected camera into a second set of eyes for a car, enabling driver-assistance and AR navigation features.
Share on XCustom license (see repository)
Overview
The Mapbox Vision AR SDK for Android uses efficient neural networks that run on the phone or embedded device, so a connected camera can act as a second set of eyes for a car. It is split into VisionCore, VisionSDK and VisionAR. VisionCore holds the machine learning models, VisionSDK wraps it in Kotlin, and VisionAR handles the augmented reality display.
Key features
- Augmented reality navigation with turn-by-turn directions
- Detection of vehicles, pedestrians, road signs and traffic lights
- Segmentation of the road scene into 14 classes
- Distance detection to the lead vehicle
Best for
Android developers building driver-assistance or AR navigation features that process camera imagery on the device.
- Upstream
- mapbox/mapbox-vision-ar-android
- Fork on GitHub
- Guo-astro/mapbox-vision-ar-android
- Upstream stars
- 14
- Category
- Maps and geospatial
- License
- Custom license (see repository)
- Forked
- 2020-12-04
- Sync status
- In syncLast synced 2026-09-29
More in Maps and geospatial
An open source command line toolkit that turns drone, balloon or kite photos into maps, point clouds, 3D models and elevation models. Used for aerial mapping and surveying.
Forked 2026-06-09Last synced 2026-09-29License: AGPL-3.0Maps and geospatialGitHub
A mobile web prototype for map lighting challenges. An organizer picks a real area and shares a link, and participants light up map cells by walking around it.
Forked 2026-06-09Last synced 2026-09-29No license declaredMaps and geospatialGitHub
A web app that generates a 3D map of real cities with building and road data, built with React Three Fiber. Maps can be exported as GLB files.
Forked 2026-05-09Last synced 2026-09-29License: MITMaps and geospatialGitHub
An open-source library for showing interactive vector tile maps in iOS, Android and desktop apps. It uses the GPU for fast rendering and is used to add maps to native applications.
Forked 2025-12-31Last synced 2026-09-29License: BSD-2-ClauseMaps and geospatialGitHub