Definition: ARCore is Google's augmented reality framework for Android devices. With it, developers create AR applications that place virtual content precisely in the real environment.
The platform uses the camera, motion sensors and processors of Android smartphones and tablets. It finds the position of the device in space, analyzes the environment and shows 3D models with the correct perspective. ARCore uses SLAM technology for markerless tracking without prepared reference points.
For B2B companies in mechanical engineering, ARCore gives access to more than 70% of smartphone users worldwide – much more than iOS. This gives a wide reach to customers, partners and service technicians, without additional hardware investments.
What Is ARCore?
Since 2017, ARCore has been Google's standard framework for augmented reality on Android. It hides the complex technology of camera tracking, 3D rendering and sensor fusion behind a developer-friendly API. Developers use this API in Java, Kotlin or Unity/Unreal Engine.
Technical basics: ARCore combines data from several sources:
- The camera captures visual information about the environment
- The IMU (inertial measurement unit: accelerometer, gyroscope) measures the movements of the device
- Motion tracking fuses sensor data for precise position tracking
- Computer vision identifies features, planes and objects in the environment - OpenGL ES / Vulkan renders virtual objects in real time
- The Depth API (optional) uses ToF sensors for precise depth sensing
The result: virtual objects stay stable in their position, also when users move. They also adapt to the lighting conditions of the environment.
ARCore is part of the extended reality ecosystem. It enables augmented reality on the largest mobile platform in the world.
Core Functions of ARCore
Motion tracking and position detection
ARCore continuously finds the position and orientation of the device in space with six degrees of freedom (6DoF): translation (x, y, z) and rotation (pitch, yaw, roll). Users can move naturally around virtual objects: move closer for details, or walk around them for a 360° view.
Tracking quality depends on the device. High-end smartphones with better sensors and processors give more stable tracking than budget devices. This is one challenge of Android fragmentation.
Environmental understanding
ARCore detects horizontal planes (floors, tables, work surfaces) and vertical planes (walls, cabinets). This allows a realistic placement of 3D models: a virtual machine stands on the hall floor and does not float in the air.
Advanced features also detect inclined planes and ceilings. The more of the environment the camera captures, the better the plane detection becomes.
Light estimation
ARCore analyzes the brightness and color temperature of the environment to light virtual objects correctly. A machine in a bright production hall looks different from the same machine in dim office light. Newer versions support Environmental HDR for more photorealistic lighting.
Depth API and occlusion
Devices with time-of-flight (ToF) sensors or dual cameras support the Depth API for precise depth maps. This enables occlusion: real objects can hide virtual objects. When an employee walks in front of a virtually placed machine, the machine correctly disappears behind the person.
On devices without hardware depth sensors, ARCore uses motion stereo: it estimates depth from several camera frames taken from slightly different positions.
Augmented images and object tracking
ARCore can detect and track 2D images (logos, QR codes, product photos). It can also scan 3D objects and recognize them later. In a B2B context, a printed product flyer becomes an AR trigger: when ARCore detects it, the 3D machine appears.
Cloud Anchors
With cloud anchoring, several Android devices (and iOS devices via ARKit) can share the same AR session. Two people at different locations see the same virtual object in the same place. This is important for remote collaboration between sites.
Geospatial API
ARCore can use GPS, compass and Google Street View data for location-accurate AR experiences outdoors. This is relevant for construction site visualization, outdoor facility planning or logistics applications.
ARCore for B2B and Mechanical Engineering
Product visualization at the customer site
Sales teams show machines and systems at full size directly at the customer site. Space requirements, accessibility and integration into existing lines become visible before the physical machine is delivered. Customers use their own Android devices – no additional hardware is necessary.
The reach of Android means that more potential customers have compatible devices. Especially in emerging markets, Android dominates with a market share of >80%.
Trade show presentations without physical exhibits
Tablets at the booth show the full product portfolio in AR. Visitors place virtual machines in front of them and explore details. There is no logistics effort for heavy exhibits, but visitors still get an immersive impression of the products.
Service and maintenance
Service technicians on site use Android smartphones and tablets for AR-supported maintenance. AR overlays show maintenance steps, component names or warnings directly on the real machine. Remote experts see the same image and can draw virtual marks to show which components to check.
Android devices are often less expensive than iOS devices. This is relevant when you equip larger service teams. Ruggedized Android devices for harsh industrial environments are available.
Training and onboarding
New employees learn to operate machines with AR instructions on tablets. Virtual hints point to controls, explain processes and warn about danger zones. Training becomes more interactive and more practical.
Factory planning and layout optimization
New machine layouts are projected virtually into existing production halls. You can check walkways, material flow and ergonomics at a 1:1 scale before you make investments.
Quality control and inspection
AR overlays show the target dimensions over real components. Deviations become visible immediately. The system records measurement and inspection reports automatically and sends them to backend systems.
ARCore vs. ARKit: Android vs. iOS
ARCore is Google's answer to Apple's ARKit. Both frameworks have similar core functions, but they are very different in reach and device quality:
ARCore (Android):
- Wide device support across different manufacturers (Samsung, Xiaomi, Oppo, etc.)
- Largest worldwide reach (~70% of smartphones, >80% in emerging markets)
- Hardware quality varies by device manufacturer and price range
- Fragmentation: different Android versions, sensors, screens
- Lower-cost entry-level devices are available (from ~€200)
- Ruggedized industrial devices are available
- Integration with the Google ecosystem (Maps, Lens, Cloud)
ARKit (iOS):
- Only for iPhone and iPad
- Close hardware and software integration through the Apple ecosystem
- Higher device quality and more uniform performance
- LiDAR support in premium models
- Smaller device base (~26% of smartphones worldwide)
- Higher price level (premium target group)
- Integration with the Apple ecosystem
For B2B mechanical engineering: We recommend a multi-platform approach. Vuframe Copilot supports both frameworks: the same 3D models work on iOS and Android. Customers use the devices they have.
Strategic considerations:
- Maximum reach → give priority to ARCore (more devices, more countries)
- Premium B2B → both platforms (decision-makers often have iOS, technicians Android)
- Consistent quality → prefer ARKit (less fragmentation)
- Cost efficiency → ARCore (less expensive devices for service teams)
- Industrial environments → ARCore (ruggedized devices are available)
ARCore Versions and Development
Google releases ARCore updates regularly:
ARCore 1.0 (2018) - Motion Tracking, Environmental Understanding, Light Estimation
ARCore 1.2 (2018) - Cloud Anchors, Augmented Images, Vertical Plane Detection
ARCore 1.7 (2019) - Augmented Faces, Environmental HDR
ARCore 1.11 (2019) - Depth API (ToF support)
ARCore 1.15 (2020) - Recording and Playback API
ARCore 1.20 (2021) - Raw Depth API, Instant Placement
ARCore 1.30 (2022) - Geospatial API
ARCore 1.40 (2023) - Semantics API, Scene Semantics
Newer features need newer Android versions and sometimes specific hardware (ToF sensors for the Depth API). B2B applications should require at least Android 7.0 (API level 24) for a solid set of baseline features.
Device Compatibility and Fragmentation
Devices that support ARCore: Google keeps a list of ARCore-certified devices. Currently (2024), there are more than 400 million active devices worldwide. Not all Android devices support ARCore. They need at least Android 7.0 and specific sensor configurations.
Fragmentation challenges:
- Performance variance: budget devices with weak processors have less stable tracking
- Sensor quality: low-cost IMUs give inaccurate data
- Camera quality: poor cameras make feature detection more difficult
- Screen sizes: the UI must work on displays from 4 to 12 inches
- Android versions: newer features are only available on newer OS versions
Best practices for fragmentation:
- Performance profiles for different device classes
- Graceful degradation (turn off features on weak devices)
- Extensive testing on a representative device base
- Communicate the minimum requirements clearly
Technical Integration and Development
Development environments: - Android Studio (official IDE for Android) - Java or Kotlin for native development - Unity or Unreal Engine for cross-platform development - WebXR for browser-based AR (experimental)
3D asset formats: - glTF/GLB (recommended for web compatibility) - OBJ, FBX (classic formats) - Filament scene format (.filamat) for optimized performance
CAD-to-ARCore workflow:
1. Export the CAD data (STEP, OBJ, FBX)
2. Optimize the 3D model (polygon reduction, texture baking)
3. Convert to glTF/GLB
4. Integrate into the ARCore app (Filament, Sceneform or a custom renderer)
5. Test on the target devices (different manufacturers and price ranges)
Modern pipelines automate steps 2–4. After you convert a model once, it also works for ARKit, web AR and other platforms.
Performance and Limitations
Performance factors:
- Polygon count: 30,000–150,000 polygons for main assets, depending on the device class
- Texture resolution: 1K–2K textures for mobile, compressed (ETC2, ASTC)
- Draw calls: minimize them with mesh batching
- Frame rate: 30–60 FPS, depending on the device
Limitations: - Tracking quality depends strongly on the hardware - Performance varies between devices (high-end vs. budget) - Limited battery life during longer AR use (20–60 minutes, depending on the device) - Overheating with complex scenes, especially on budget devices - No precision measurements (cm accuracy, not mm) - Outdoor tracking is difficult in direct sunlight or in environments with little texture
Best practices: - Progressive loading of complex models - LOD (level of detail) for large scenes - Device profiling and adaptive quality - Thermal management (reduce performance when the device overheats) - Regular tests on budget devices (not only flagships)
Frequently Asked Questions (FAQ)
Which Android devices support ARCore?
ARCore needs at least Android 7.0 (Nougat) and specific sensor configurations. Google keeps an official list of supported devices: more than 400 million active devices worldwide from Samsung, Xiaomi, Oppo, OnePlus, Google Pixel and many other manufacturers. Important: not all Android devices support ARCore – especially older and very low-cost models do not. Our recommendation for B2B: mid-range devices or better (from ~€300) for stable tracking. Check the list at: https://developers.google.com/ar/devices
What is the main advantage of ARCore over ARKit?
Reach. ARCore runs on more than 70% of smartphones worldwide – much more than iOS (~26%). Android dominates especially in emerging markets and with price-sensitive target groups. For B2B companies with international customers or large service teams, this means: customers and technicians are more likely to already own compatible devices. Other advantages: less expensive hardware options, ruggedized industrial devices, Google Cloud integration. Disadvantage: the varying hardware quality needs more testing effort than iOS.
Can you use existing CAD models for ARCore?
Yes. You can convert CAD data from SOLIDWORKS, CATIA and Siemens NX for ARCore. Workflow: export as STEP/OBJ → polygon reduction (more important than for ARKit because of weaker devices) → texture optimization → conversion to glTF/GLB. Important: you must optimize the models for different device classes. Budget devices need lower polygon counts than flagships. Best practice: implement an LOD system. The same optimized models also work for ARKit, web AR and VR.
How do I handle Android fragmentation?
Fragmentation is a challenge, but you can manage it: 1) Define minimum requirements (for example Android 8+, mid-range devices), 2) Test on a representative device base (Samsung flagships, Xiaomi mid-range, budget devices), 3) Implement performance profiles (adjust the quality automatically), 4) Use graceful degradation (turn off features on weak devices), 5) Communicate the device requirements clearly to users. Alternative: SaaS solutions such as Vuframe Copilot handle fragmentation automatically and give optimized experiences across device classes.
How much does it cost to develop an ARCore app?
The costs are similar to ARKit apps. A simple app with product visualization: €15,000–40,000 (basic UI, ARCore integration, 1–3 products). An extensive solution with configurator, backend and multi-user: €50,000–150,000. Additional effort with ARCore: testing on more devices (+10–20% testing time), performance optimization for budget devices (+10–15% development time). Running costs: Google Play developer account ($25 one-time fee), hosting, updates. Alternative: SaaS such as Vuframe Copilot. CAD conversion: €2,000–8,000 per model. Wide reach and shorter sales cycles often give a return on investment (ROI) in 12–18 months.
Industry Standards and Technologies
Google ARCore documentation Official developer documentation with API reference, tutorials and best practices. Source: https://developers.google.com/ar
ARCore GitHub Open-source examples and SDKs for different platforms. Source: https://github.com/google-ar/arcore-android-sdk
OpenXR standard Cross-platform standard of the Khronos Group for AR/VR. ARCore supports it more and more. Source: https://www.khronos.org/openxr/
glTF 2.0 format Recommended 3D format for ARCore apps, optimized for mobile performance. Source: https://www.khronos.org/gltf/
Filament rendering engine Google's physically based rendering engine, optimized for ARCore and mobile devices. Source: https://google.github.io/filament/
Related Topics
AR technologies: - Augmented Reality (AR) – digital overlays in the real world - ARKit – Apple's AR framework for iOS - Markerless Tracking – AR without prepared markers - SLAM – the technology behind ARCore motion tracking - Image Tracking – AR activation through image recognition
3D technologies: - 3D Model – digital representation for AR - glTF – cross-platform 3D format - CAD-to-3D conversion – workflow from design to AR - Real-time Rendering – 3D display in real time - LOD (Level of Detail) – performance optimization
Immersive technologies: - Extended Reality (XR) – umbrella term for AR/VR/MR - Virtual Reality (VR) – full virtual immersion - Mixed Reality (MR) – virtual objects anchored in space - Mobile AR – AR on smartphones and tablets - WebAR – browser-based AR
Platforms and devices: - Android – mobile operating system for ARCore - Depth Sensors – ToF and stereo for precise tracking - Ruggedized Devices – AR devices for industrial use - Cloud Anchors – multi-user AR experiences - Geospatial AR – location-accurate AR outdoors
Areas of application: - Product Visualization – product presentation in AR - Remote Service – technical support with AR - Mobile AR – AR on smartphones - Industrial AR – AR in production environments - AR Training – training with AR instructions
