Automated Vehicle
Damage Inspection

AI-powered scanning gate captures, identifies, and inspects every vehicle passing through — detecting surface damage in real time.

9
Cameras per Gate
<18s
Full Scan Time
24/7
Continuous Operation
Scroll to explore
How It Works

End-to-End Automated Pipeline

From the moment a vehicle enters the gate to a complete damage report — fully automated, no manual intervention required.

Step 1
Movement Detection
Step 2
Plate Recognition
Step 3
Image Archival
Step 4
AI Detection
Step 5
Deduplication
Step 6
Report
The Hardware

Scanning Gate

A purpose-built metal gate with cameras, LED lighting, and weather covers — vehicles drive through at normal speed while being captured from every angle.

Drag to rotate · Scroll to zoom
Dashboard

Central Dashboard

Real-time overview of all scanning activity. Track daily vehicle counts, review the latest scan events, and monitor system health from a single view.

  • 7-day rolling event chart
  • Instant access to any scan event
  • Front & rear plate display
  • Detection count summaries
Scan Viewer

Multi-Camera Scan Viewer

Full-resolution image viewer with all 9 camera angles. Navigate through timestamps and cameras to inspect every inch of the vehicle surface.

  • Thumbnail navigation across all cameras
  • AI detection overlays with bounding boxes
  • Zoom and pan for detail inspection
  • Confidence scores per detection
Auto-View

Cinematic Scan Walkthrough

Automatic playback steps through every camera and frame, highlighting AI detections as they appear — perfect for quick review.

AI-Powered

Damage Detection

Custom AI vision models locate vehicles and analyze their surfaces for damage — scratches, dents, cracks, paint defects, and missing parts.

New Damages Detected

New Damage Alerts

When a vehicle is scanned, the system cross-references with all previous scans. Newly detected damages are flagged immediately.

  • Automatic comparison with scan history
  • Confidence scoring per finding
  • One-click jump to full scan view
Damage Comparison

Scan-to-Scan Comparison

Side-by-side view comparing the current scan with the vehicle's previous visit. Each damage is classified as new, previously known, or worsened.

  • Color-coded: new / changed / known
  • Matched by license plate across visits
  • Spatial similarity scoring algorithm
  • Summary statistics per comparison
Manual Annotation

Manual Annotation & Self-Training

Operators can draw polygon annotations directly on scan images to mark new damage types, stickers, or other elements of interest. These annotations feed back into the AI training pipeline.

  • Polygon, rectangle, and freeform tools
  • Define new detection classes on the fly
  • Approval workflow (pending / validated / rejected)
  • Controlled self-training with checkpoint rollback
Under the Hood

Technology Stack

Purpose-built from the ground up for high-throughput, real-time vehicle inspection.

AI Models

Custom-trained vision models for plate detection, vehicle detection, and damage classification. Self-training capability with checkpoint management.

Cameras

High-resolution, low-light-sensitivity cameras per gate. Full 360-degree vehicle coverage at high frame rate capture.

Real-Time Processing

High-performance state management for live progress tracking, image caching, and sub-second UI updates during scanning.

ALPR Integration

Automatic License Plate Recognition with vehicle registry validation. Independent front and rear plate reading.

Extensible Sensors

The system supports additional sensors attached to the scanning gate — height measurement, weight, temperature, and more — matched automatically with scan events.

Automated Storage

Automatic data lifecycle management with configurable retention policies and space-aware cleanup of historical scan data.