We are revolutionising road maintenance using AI and visual cameras to locate, classify, and assess pavement defects quickly and accurately — turning every inspection drive into a prioritised maintenance plan.
Why automate pavement inspection
Manual pavement inspections are slow, expensive, and inconsistent. By the time a defect is recorded, ranked, and scheduled for repair, it has often grown into a much costlier failure. Our defect detection module brings continuous, objective coverage to every street you manage.
Automated inspection
Replace hand-written reports with automated detection and reporting — reducing labour and time costs across the network.
Spot defects early
Catch hairline cracks before they become potholes. Early intervention keeps repairs small, cheap, and safe.
Prioritise objectively
Measured, repeatable defect data drives evidence-based maintenance prioritisation — not guesswork.
Live defect detection running on inspection footage — cracks and potholes flagged automatically.
How the module works
1. Detection — YOLO
A real-time YOLO detection model scans every inspection frame and instantly identifies and classifies pavement defects: longitudinal and transverse cracks, alligator cracking, potholes, edge breaks, and surface failures.
2. Measurement — UNet
A specialised UNet segmentation model accurately measures the geometry of each defect — length, width, and area — for precise quantification and severity scoring.
3. Reporting — GPT-4 Vision
A vision language model automatically drafts narrative inspection reports from the visual data, eliminating manual write-up. Defects are linked back to GPS positions and timestamps so every observation is auditable.
The automated process
Camera images go in, actionable maintenance reports come out. The result is a fast, reliable, and cost-effective pavement assessment system that fits straight into the rest of the iRoadTech digital twin — defects, signs, and kerbs all visible on the same map.
Pair it with
Smartphone roughness
Add network-wide pavement roughness measurement using everyday smartphones mounted in vehicles.
Kerb & Channel
Inspect kerbs and channels in the same drive — defects and edge condition together in one report.
Visualisation Portal
Browse defects on a live map alongside parking signs and other detected assets.