EXPLAINER · 2026

What is a digital twin and how does it save money on infrastructure projects?

TL;DR

A digital twin is a measurable, photorealistic 3D model of your real asset, built from drone imagery and delivered in the formats your design software already reads. The savings come from five places: fewer site visits, remote measurement, a permanent condition record, safer planning of access work, and clash detection against design models.

Textured photogrammetric mesh of a riveted bridge truss node from a digital twin

What a digital twin is

Strip the buzzword back and a digital twin is simple: a three-dimensional, dimensionally accurate, photographically textured model of a physical asset, current as of the day it was captured. Unlike a design model, it records what is standing on site, down to individual members, connections and surface condition.

The level-of-development language from BIM practice is useful here. An LOD 400 twin carries fabrication-level detail: your engineers can take a dimension off a gusset plate in the model to a relative accuracy of up to 3–5 mm, without anyone putting a tape on it, because the model was validated against surveyed ground control when it was built.

LOD and accuracy are different things on a captured model

The level-of-development scale was written for models that are drawn. LOD 300 is coordinated design geometry; LOD 400 is fabrication-level detail, with member sizes, connections and assembly shown. A captured twin can carry that richness without anyone modelling it, because the camera records the plates, bolts and brackets the flight plan was designed to resolve. What the LOD label does not state is how closely that geometry sits to reality.

That second property is accuracy, and on a captured model it comes from the survey workflow underneath the imagery: RTK positioning against a base station, ground control measured independently, and check points withheld from the bundle adjustment so the finished model is tested against marks it never used. Two models can both look like LOD 400 and differ by centimetres. The useful specification for a captured twin is two numbers side by side, the detail you expect to see and the accuracy class the model will be held to, with the residuals reported in a Survey Quality Report. If a quote names an LOD and no accuracy class, ask for the second number.

How one gets made

The workflow has three stages. First, capture: an enterprise RTK drone platform flies a designed pattern around the asset, recording hundreds or thousands of overlapping images, each position-stamped against a base station to a horizontal accuracy of up to 1 cm + 1 ppm. Second, processing: professional photogrammetry software matches features across the images and reconstructs them into a point cloud and a textured mesh, which is then constrained by surveyed control points and tested against check points withheld from the adjustment. Third, delivery: the validated model is cut into the formats your team works in, with the Survey Quality Report attached.

For a single structure, capture is typically one site visit. Nobody climbs anything, no plant is hired, and the asset keeps operating underneath. The full method sits on our photogrammetry and digital twins page.

Where the money comes back

Cutting site visits is the first and largest saving. Every question that can be answered from the model (a dimension, a clearance, a condition query, a constructability check) is a trip, a permit, and a slice of risk exposure that never happens. On a regional asset, avoiding a single mobilisation can save more than the twin costs.

Remote measurement is the second. Once the twin exists, anyone with access can measure anything on the asset, at any time, without a booking. Design teams stop working from decade-old drawings that no longer match reality.

The condition record is the third. A twin is a timestamp: the asset exactly as it stood on capture day. Repeat the capture next year and change becomes measurable (settlement, corrosion growth, works progress) rather than argued about from mismatched photographs.

Safer, cheaper access planning is the fourth. Scaffold and rope-access designers can plan against the real geometry in the model, so crews arrive with a design that fits first time.

Clash detection is the fifth. Drop the twin into the same environment as the as-designed CAD and mismatches between drawing and reality surface on a screen, before they surface as a variation claim.

Formats and software: it has to fit your stack

A twin that won't open in your software goes unused. Deliverables should arrive as textured mesh (OBJ or FBX), point cloud (LAS or RCP), and georeferenced imagery: formats that drop directly into Civil 3D, 12d, Revit, OpenRoads and Trimble Business Center without a conversion project. If a provider cannot tell you exactly which formats you will receive and what software they open in, keep shopping.

Lifecycle management: keeping the twin current

The highest-value twins are the ones that get updated. A capture at handover establishes the baseline; scheduled recaptures at inspection intervals build a condition history that maintenance can be planned from. Because our capture runs on fixed, repeatable parameters, each cycle is directly comparable with the last — the same faces, the same resolution, years apart.

For Brisbane and South East Queensland assets, that program can start with a single structure; our Brisbane operations page covers mobilisation, airspace and what a first capture involves.

NEXT STEP

Talk to the pilot who flies the job

Send the asset and the deliverable you need. A scoped method and quote come back within one business day.