Model to Field: Layout, AR & Reality Capture

Getting the model *onto* the site takes two forms. Digital layout replaces the tape measure and the string line: a robotic total station (or similar positioning tool) takes the exact coordinates of points from the model — where a wall starts, where an anchor bolt goes, where a penetration is needed — and marks them on the real site to millimetre accuracy. The setting-out that used to be manual, slow and error-prone becomes a direct transfer from the coordinated model to the ground, so what gets built sits exactly where the model said.

Augmented reality goes further, overlaying the model onto the worker's actual view — through a tablet, a phone or a headset — so they can *see* the design in place: this duct runs here, at this height; this is where the services sit above the ceiling before it is closed. AR turns the model from something you consult on a screen into something you see superimposed on the work in front of you, catching 'that's not where I thought it went' before it is built wrong. Both technologies share a precondition worth stating plainly: they are only as good as the model behind them. Lay out precisely from a wrong model and you build the wrong thing, precisely.

The reverse direction may matter even more. Reality capture records the physical site as it actually is — most powerfully through laser scanning, which produces a dense point cloud of millions of measured points, and through photogrammetry, which builds 3D geometry from photographs. The result is a precise, measured record of what is really there — the existing building before you start, or the new work as it is actually built.

That record does two crucial jobs. First, scan-to-BIM: capturing an existing or heritage building as a point cloud and modelling from it, so renovation and retrofit start from reality rather than guesswork (this is the basis of HBIM, Module 8). Second, as-built verification and progress monitoring: scanning the new construction and comparing it against the design model, to check that what was built matches what was designed — and to measure, objectively, how much has been done. This is the honest counterweight to every optimistic claim on a construction site: not 'the drawings say it should be here', but 'the scan shows where it actually is'. Reality capture is how the model is held accountable to the ground truth.

Put the two directions together and you have a closed loop between model and reality: the model drives the layout and guides the work (model to field), and reality capture checks the result and feeds the true as-built state back (field to model). That loop is how a project keeps its model and its building in agreement instead of letting them silently drift apart — and it is what makes the honest 5D payment of lesson 7.2 and the trustworthy handover of the next module possible, because the as-built model that flows into operation is verified against a scan, not just assumed.

The honesty, one last time for this module, is that these are powerful aids, not autopilot. Digital layout and AR are only as truthful as the model; a confident overlay of a wrong model is confidently wrong. Reality capture produces enormous, precise data, but a point cloud is not a model — turning it into useful information takes skill and judgement, and comparing scan to design still needs a human to decide what a deviation means. The field is messy, and no technology removes the need for care. But used well, this loop is the fitting close to BIM in construction: it takes the model out of the office and onto the ground, and — just as importantly — it drags the ground truth back to the model, so that everything downstream is built on what is *actually there*, not what someone hoped was there. That is the discipline the whole course has been arguing for, made physical on a construction site.