Scan-to-BIM with Drone Data: Workflow and Accuracy

Table of Contents

What Scan-to-BIM actually involves

Scan-to-BIM is the process of turning 3D scan data – from terrestrial laser scanning, drone photogrammetry, or drone LiDAR – into a Building Information Model in Revit, ArchiCAD, or Bentley. The BIM model is a geometrically accurate, intelligent 3D representation of what’s actually there, useful for design, analysis, and facilities management.

Drone data has become a big part of Scan-to-BIM workflows, particularly for building exteriors, roofs, site context, and large infrastructure that you can’t practically survey from ground level.

What the drone contributes

Roofs and upper facades

Terrestrial scanners are great for interiors but they can’t see roofs, upper facades, or anything without clear line-of-sight from ground level. Drones fill that gap:

• Complete roof geometry – ridges, valleys, hips, eaves, parapets, roof plant, access structures
• Upper facade detail – windows, cladding joints, structural elements, defects on upper storeys
• Full building envelope that combines with interior scan data for a complete model

Site context

BIM models need accurate surroundings for coordination, planning compliance, and presentations. A drone topographical survey gives you:

• Existing ground levels and terrain
• Adjacent buildings, roads, infrastructure
• Vegetation, boundaries, site features
• Utility corridors and access

Large infrastructure

For infrastructure BIM – bridges, retaining structures, embankments, road corridors – drones capture external geometry far more efficiently than terrestrial scanning. A bridge that would need multiple scanner setups and possible lane closures can be captured in one flight.

The workflow

Stage 1: Data capture

We plan specifically for Scan-to-BIM:

• Grid plus oblique (45 degree) flight pattern to cover all facades and roof surfaces
• Low altitude, high overlap for sub-centimetre resolution on architectural detail
• GCPs surveyed in the project coordinate system so drone data aligns with terrestrial scan data

Stage 2: Point cloud generation

Photogrammetric processing produces a dense, coloured point cloud. For LiDAR, the cloud comes directly from the scanner data. Either way, accuracy gets verified against check points – typically plus or minus 2 to 5 cm.

Stage 3: Registration and merging

If the project has both terrestrial scanning (interiors) and drone survey (exteriors/roof), the datasets need to be registered – aligned into one coordinate system. We do this through:

• Shared survey control points visible in both datasets
• Cloud-to-cloud registration using overlapping areas (usually facades captured by both)
• Quality checking the merged result

Stage 4: BIM modelling

The registered point cloud goes into Revit (or whichever platform) as a reference. Modellers trace geometry from the cloud to create BIM objects – walls, floors, roofs, columns, windows, MEP. Detail depends on the LOD specified:

• LOD 200: Approximate geometry – feasibility and concept
• LOD 300: Accurate geometry with correct dimensions – detailed design coordination
• LOD 350: Accurate geometry with connections – construction coordination

Accuracy

Drone data

• Photogrammetry point cloud: plus or minus 2 to 5 cm absolute (with GCPs)
• LiDAR point cloud: plus or minus 2 to 3 cm
• Surface detail: features 5 cm and larger are typically resolvable

BIM model

• LOD 200: plus or minus 5 to 10 cm – representative but approximate
• LOD 300: plus or minus 2 to 5 cm – accurate as-built dimensions
• LOD 350: plus or minus 1 to 3 cm – supports clash detection

For most renovation and extension projects, LOD 300 from drone data is appropriate and achievable. More in our accuracy guide.

Where drone data is essential

• Heritage and conservation: complex roofs, inaccessible upper elements, facade detail on protected structures
• Large commercial buildings: roof plant, multiple roof levels, extensive site context
• Infrastructure: bridges, retaining walls, structures with limited ground access
• Campus sites: where relationships between buildings matter as much as individual geometry

The level of detail (LOD) question is important for scan-to-BIM projects because it directly affects cost and timeline. LOD 200 – approximate geometry with basic dimensional accuracy – can be produced relatively quickly from drone data. LOD 300 – precise geometry suitable for coordination – requires more careful modelling and is where most Irish architectural and engineering projects sit. LOD 400 and above typically needs supplementary data beyond what a drone captures, such as internal measurements and material sampling.

From our experience, drone-captured data combined with some targeted internal measurements (which we can also provide using handheld scanners) delivers a solid LOD 300 model that most design teams can work with. The key is defining the LOD requirement before the survey so we capture the right data density and angles for the modelling team.

What we deliver

Cleaned, classified point clouds in formats for Revit (via ReCap), ArchiCAD, and Bentley. For combined terrestrial plus aerial projects, we coordinate with your scanning team to ensure everything merges cleanly.

Get in touch about your Scan-to-BIM requirements, or see our 3D modelling service.

The Irish BIM Mandate and why Scan-to-BIM is now a public-sector requirement

The Irish public-sector BIM mandate is now a firm timetable, not an aspiration, and every engineering consultancy working on state-funded projects needs to understand where the thresholds sit.

• January 2024: BIM methodology required on all public works projects valued at €100 million or more, applied to Design Teams.
• June 2025: threshold lowered to €20 million or more, and extended to cover both Design Teams and Contractors.
• End of 2027: BIM methodology required on all Irish public works projects regardless of capital value, for Design Teams, Contractors and the full project lifecycle.

The mandate is anchored in the ISO 19650 series of standards, which defines the information management framework for the whole asset lifecycle. The NSAI (National Standards Authority of Ireland) offers ISO 19650-2 certification for organisations delivering BIM projects in Ireland, and the Irish National Annex provides localised guidance on how ISO 19650 is applied to Irish public-sector procurement. For engineering consultancies, certification against ISO 19650-2 is increasingly treated as a pass or fail line on public-sector tender responses.

The governance document that ties this together is the Capital Works Management Framework (CWMF), maintained by the Department of Housing, Local Government and Heritage (DHLGH). The CWMF is the contract environment, the procurement guidance and the deliverable schedule for Irish public works. Drone survey data, specifically Scan-to-BIM workflows, slot into the topographic, as-built and existing-conditions components of the CWMF deliverable schedule.

LOD 400 and LOD 500: where drone data stops and where it keeps going

Most Scan-to-BIM articles stop at LOD 350. That is fine for design coordination, but it misses the two levels where drone data starts to drive real value for asset owners and facilities management teams.

• LOD 400 (fabrication-level detail): model elements are represented with the size, shape, location, quantity and orientation they will actually have when fabricated and installed. Drone reality capture supports LOD 400 on structural steel retrofits, facade refurbishments and heritage restoration work, where the captured geometry directly informs the fabrication model.
• LOD 500 (as-maintained, facilities management): the model represents the asset as it is, with verified field conditions. This is where drone reality capture becomes a lifecycle asset, not just a design deliverable. Asset owners who commission an as-built drone capture at project handover and then recapture at major maintenance intervals build up a time-series record of their asset that traditional survey simply cannot match for cost or completeness.

LOD 500 is the level at which Scan-to-BIM stops being a design workflow and starts being an operational workflow. For lifecycle clients with large property portfolios, OPW, local authorities, Uisce Éireann, Coillte, ESB Networks and major commercial landlords, this is where the business case for drone survey stops being a cost and starts being an asset management tool.

LOD definitions and what drone data can realistically support

The BIM world uses Level of Development (LOD) to describe how detailed a modelled element is. The AIA and BIMForum specifications define LOD 100 (conceptual), LOD 200 (approximate geometry), LOD 300 (precise geometry), LOD 350 (geometry plus interfaces), and LOD 400 (fabrication detail). Drone-captured reality data does not map cleanly to a single LOD because it captures what is physically there, and the modeller decides what to do with it.

In our experience, drone data reliably supports the following:

• LOD 200 in two days. A 20 hectare site can be flown, processed, and turned into an approximate massing model by a modeller in two working days. Useful for early concept stages and site appraisal.
• LOD 300 in five to ten days. Precise geometry for roofs, facades, and site context. This is where drone data genuinely outperforms terrestrial scanning on large or high-level elements.
• LOD 350 on request. Interface resolution (where one element meets another) typically requires terrestrial scanning at the connection points, but drone data supplies the bulk geometry and saves days of TLS work.
• LOD 400 is not a drone workflow. Fabrication-level detail belongs with measured surveys and TLS, not drone capture.

Tolerances by building element

Different building elements have very different tolerance expectations, and a Scan-to-BIM specification should spell these out at the start of the project rather than leaving them to be argued over at handover. Typical working tolerances we see on Irish projects are:

• Facades and curtain walling: ±15 to 25 mm – drone photogrammetry at short range comfortably meets this.
• Roof planes and parapets: ±25 to 40 mm – drone is the preferred capture method because no scaffold is needed.
• Structural columns and beams (exposed): ±5 to 10 mm – terrestrial scanning territory, not drone.
• Heritage masonry mouldings: ±3 to 5 mm – close-range photogrammetry or TLS.
• Site topography (landscaping, drainage): ±20 to 30 mm – drone photogrammetry or LiDAR as standard.

Getting these numbers agreed early, written into the Employer’s Information Requirements, and baked into the BIM Execution Plan prevents the most common argument on Scan-to-BIM handovers: the modeller delivered to 30 mm, the structural engineer expected 10 mm, and nobody has budget for a reshoot.

Irish heritage and protected structures

A meaningful share of our Scan-to-BIM work in Ireland is on protected structures – buildings on the Record of Protected Structures maintained by each local authority, or National Monuments under the care of the OPW. These projects have two specific considerations worth flagging.

First, the condition survey typically needs to resolve individual stone courses, pointing width, visible cracks, and vegetation growth on the fabric. This requires drone imagery at 1 to 2 mm ground sample distance, which in turn requires flight altitudes of 5 to 8 metres from the facade. That is low enough to need extra care with wind, public access, and nesting birds, and we plan around swift and swallow nesting season by default.

Second, the Conservation Officer often wants the output as annotated elevations and high-resolution orthophotos alongside the BIM model, not in place of it. We therefore produce both the textured mesh for the Scan-to-BIM workflow and a set of rectified facade orthophotos at 1:20 or 1:50 that the conservation team can mark up directly. For context on how this sits alongside our wider heritage work, see our heritage recording service.

Frequently Asked Questions