Vision AI Platform
Computer vision platform powering food recognition, drive-thru optimization, and operational intelligence across restaurant operations.
Overview
The Vision AI Platform provides real-time computer vision capabilities for restaurant operations. It uses state-of-the-art models (YOLOv8, custom CNNs) to recognize food items, detect drive-thru lines, predict wait times, and provide operational insights from surveillance cameras.
Capabilities
| Capability | Description | Page |
|---|---|---|
| Food Item Recognition | KDS visual verification of order accuracy at the expo station | Image Analysis |
| Drive-Thru Detection | Real-time queue monitoring and wait time prediction | Drive-Thru Detection |
| Operational Intelligence | Surveillance camera AI for dining, kitchen, and safety | Operational Intelligence |
| Visual Facility Mapping | AI-powered floor plan extraction from photos and CAD files | Facility Mapping |
| Visual Inventory Counting | Camera and mobile-based inventory counting | Inventory Counting |
| Custom Task Builder | No-code monitoring rules, verification tasks, and alerts | Custom Tasks |
| Edge Deployment | On-premise hardware, cameras, and cloud sync | Edge Deployment |
| API Reference | REST endpoints for cameras, drive-thru, orders, and occupancy | API Reference |
| Manager Dashboard | Step-by-step guide for monitoring and acting on Vision AI insights | Manager Guide |
Architecture
All vision AI processing runs on edge devices at each location. Only metadata and aggregated metrics sync to the cloud -- video footage never leaves the premises.
Camera Feeds --> Edge Device (Jetson) --> AI Models (YOLOv8) --> Detection Events
|
v
Cloud Sync (metadata only)
|
v
Cockpit Dashboards
Related Documentation
- ACP AI Router - AI infrastructure
- Edge Infrastructure - OlympusEdge
- Inventory API - Inventory management