Efficiency
Turn large archives of tunnel images into structured reports without manual surveying.

A FAST and specialized tunnel inspector.
See every defect before the next inspection.
Inspection, automated end to end.
DART is the AI vision system behind automated tunnel and civil-structure inspection. It recognises and classifies defects from ordinary images, validates field surveys, and returns objective, repeatable results to the people who maintain the network.
Turn large archives of tunnel images into structured reports without manual surveying.
A single vision model applies the same criteria everywhere, cutting classification mistakes.
Results are graded consistently over time, making inspections comparable and auditable.
Upload an image archive, set the parameters that matter, and launch the analysis. DART returns an interactive dashboard of the inspected asset.
Drop the image archive as a .zip. No special capture rig or positioning required.
Set the inspected length and the detection thresholds before the run.
The vision model processes every frame and classifies the defects it finds.
A summary chart, per-segment frequency, defect-type distribution and a downloadable table.
Read the tunnel as a whole, then drill into the segments and defect types that matter.


A single chart of defect density along the entire inspected length.

A histogram showing how defects cluster from one segment to the next.

The breakdown of defect types, with the underlying table ready to download as CSV.
From network operators to inspection teams, wherever structural health has to be measured, recorded and compared over time.
DART is developed with research excellence and industry leaders across infrastructure and civil engineering.


Recognition
85%
F1 macro
0.77
Training photos
~10,000
Reach us to bring DART into your next inspection programme.