Home / Demo
PIVAI Atlas · Precision survey demonstration

One drive. Every crack, measured.

In October 2026 we surveyed a two-lane asphalt road in South Carolina with PIVAI Atlas to show what a single drive-through can deliver.

Roadway814mof centreline surveyed in one drive.
Cracks mapped7,457m1.09 m of crack per m² on average.
Imagery6,831m²of asphalt at 2 mm per pixel, plus a 5 mm 3-D surface.
Condition60indicative crack condition index: Fair.
How the data was captured

Survey-grade, with no lane closures.

Atlas was mounted on an ordinary vehicle and driven at 20 mph in normal traffic. Three rear-facing industrial cameras, fanned to overlap across the lane, recorded the pavement while a GNSS/INS logged position and orientation for every frame.

The images were combined into one seamless 2 mm image of the road and a 5 mm elevation model. PIVAI's AI then found every visible crack, measured it and classified it.

Map sheet: A, the 2 mm imagery with detected cracks coloured by type along the road; B, crack density on a 0.5 m grid with 25 m sample-unit condition ratings
A map-book sheet: (A) the 2 mm imagery with detected cracks by type; (B) crack density on a 0.5 m grid, with pattern areas and a condition rating for every 25 m.
What we found

Not just where it's cracked: how, and how badly.

Each crack is classified by the pattern it forms with its neighbours, following the distress definitions in ASTM D6433.

Three examples: isolated cracks in red, block-pattern cracking in blue and alligator-pattern cracking in purple, each shown on the 2 mm imagery and with AI detection
■ Isolated longitudinal / transverse   ■ Block pattern   ■ Alligator (fatigue) pattern · scale bar 1 m
FINDING 01

Fair overall, with clear hot spots

Of 33 sample units (25 m each): 3 Good, 10 Satisfactory, 4 Fair, 14 Poor and 2 Very Poor. The worst 250 m stretch accounts for most of the block and alligator cracking.

FINDING 02

The joint, not the wheel path

Cracking peaks within about 1 m of the centre line, pointing to the centre-line construction joint as the main source of deterioration rather than traffic fatigue.

FINDING 03

Mostly early-stage cracking

69% of crack length is narrower than 10 mm (low severity), and transverse cracks dominate (44%): the profile of ageing asphalt that is still in the preservation window.

Ready for maintenance planning

Quantities that go straight into an estimate.

Because every crack is measured, the survey produces scoping quantities directly, not just a score.

  • Crack sealing1,050 m of isolated cracks ≥ 10 mm
  • Crack filling2,052 m of isolated cracks < 10 mm
  • Joint sealing814 m of centre-line joint, full length
  • Block cracking1,294 m² (18.9% of the asphalt)
  • Alligator cracking370 m² (5.4% of the asphalt)
  • RehabilitationOne candidate section of 1,960 m²
Deliverables

What the client receives.

Everything shares one coordinate reference, so imagery, crack maps and tables overlay exactly in any GIS or CAD package.

Condition report

Summary, ratings, worst sections, maintenance quantities, method and limitations.

Map book, 11 × 17 in

Every detected crack drawn on the 2 mm imagery, with statistics for each sheet.

GIS & CAD data

2 mm imagery and 5 mm elevation GeoTIFFs, classified crack map, GeoJSON and CSV.

Google Earth package

Imagery, cracks and road sections coloured by condition in a single KMZ.

Argus video

Roadside asset detection video recorded with PIVAI Argus on nearby roads.

Next steps

Network-wide surveys, repeat surveys for change detection, and data for pavement management systems.

Indicative results from a demonstration survey. The crack condition index is a crack-based proxy for PCI following ASTM D6433 methods in simplified form.

Get in touch

Want this for your roads?

We run demonstration surveys for agencies, airports and consultants. Tell us about your network.