Industrial Mobile AR Application Development
At an oil refinery, a technician replaces a valve. Every hour of downtime costs the enterprise about $9k–13k in savings. Documentation is a 400-page PDF, schematics in AutoCAD on a server, instructions in a folder. Searching for the right information takes up to 4 hours. An AR app on a tablet shows an arrow to the correct valve, displays bolt torque, and step-by-step instructions overlaid on the real equipment. Replacement time drops by 60%.
Such solutions run on Vuforia, Scope AR, Augmentir. We build similar systems, adapting them to corporate ERP and the specifics of a given production facility. Contact us for a preliminary assessment—we'll evaluate your project.
How to Recognize Industrial Equipment Without Markers?
Standard image tracking on a flat picture doesn't work in a workshop: equipment is three-dimensional, metal surfaces glare, lighting is unstable, and smoke may be present. Let's compare three approaches:
| Method | Accuracy | Deployment Complexity | Environmental Dependency | Cost |
|---|---|---|---|---|
| Marker-based (ArUco/QR) | High | Low (stick markers) | None (markers always visible) | Low |
| Markerless object recognition (ARKit/ARCore) | Medium–High | Medium (scan object) | Sensitive to lighting and texture | Medium |
| Spatial anchors (Azure/ARCore Cloud) | High | High (cloud infrastructure) | Minimal | High |
Marker-based — we use ARImageTrackingConfiguration (ARKit) or AugmentedImageDatabase (ARCore). Reliable, independent of the object's appearance. Downside: markers must be physically placed and may get dirty.
Markerless object recognition — ARKit provides ARObjectScanningConfiguration for scanning 3D objects into .arobject format. Works for compact equipment with texture. For large objects, we train a model via Create ML or use Vuforia Model Targets SDK.
Spatial anchors — position the scene once and save the anchor to the cloud. On restart or device change, the position is restored. Suitable for large objects: turbines, tanks, machine tools.
Why is AR Useless Without ERP Integration?
Industrial AR without connection to production systems is an expensive toy. The real value is in contextual data:
- Scanning equipment pulls up the current maintenance order, repair history, and required spare parts.
- Overlays of real sensor readings (temperature, pressure, vibration) on physical equipment via OPC-UA or MQTT.
- Digital equipment passport: users see an AR layer with ERP data directly over the machine.
Integrations via REST or GraphQL, authentication via corporate SSO (SAML 2.0, Azure AD). Offline mode is critical: a production shop may not have stable Wi-Fi; all instructions are cached locally using Core Data or Room.
What Technologies Do We Use?
| Component | Technology |
|---|---|
| Object recognition | ARKit (iOS), ARCore (Android), Vuforia |
| 3D rendering | SceneKit, RealityKit, Unity |
| Data caching | Core Data, Room |
| ERP integration | REST, SOAP, GraphQL |
| Authorization | SAML 2.0, Azure AD |
What's Included in a Pilot Project?
- Equipment and process audit — select 10–20 units for the first AR instructions.
- Choose recognition method — based on audit results: markers, object tracking, or spatial anchors.
- Develop a prototype — for one equipment type, linked to real data.
- Integrate with ERP/SCADA — pull current orders and sensor readings.
- Create an authoring tool — so technicians can create new instructions without programmers.
- Pilot deployment — test in one workshop for 2 weeks.
Result: a working prototype, API documentation, and training for two employees. Timeline: 8–12 weeks. Cost is calculated individually — call or email us, we'll evaluate your project within a day.
Steps for Implementing an AR Solution on the Production Floor
- Pre-project survey (3–5 days) — engineer visit to site, photography, condition analysis.
- Stack selection: AWS + ARCore/ARKit + ERP adapter.
- MVP development (4–6 weeks) — basic AR instruction for one assembly.
- Testing on real equipment with technician participation.
- Refinement and scaling to remaining units.
- Commissioning and staff training.
- Support (12-month warranty on architectural solutions).
Why Choose Us
- Over 10 years of experience in developing AR for industry.
- Delivered 40+ projects for machine-building, oil & gas, and energy enterprises.
- Proprietary authoring tool platform — cuts time to create new instructions by half.
- Warranty on all architectural solutions — 12 months.
- Work with any ERP (SAP, 1C, Oracle) and SCADA.
Order a pilot project — get a working prototype in 8–12 weeks. Contact us by phone or through the form on our website. Get a consultation for your project — we'll assess complexity and propose the optimal solution.
ARKit from Apple — official documentation.