Point Cloud to BIM Services — Complete Guide for AEC Professionals

[edit] 1. Overview

Point cloud to BIM is the process of converting 3D laser scan data — a dense collection of measured XYZ coordinates captured by a terrestrial or mobile laser scanner — into an intelligent, parametric Building Information Model. The output is not a rendered visual or a 2D drawing set. It is a structured, data-rich model built within authoring software such as Autodesk Revit, in which each element (wall, column, duct, beam) carries geometric and non-geometric attributes.
The process is distinct from Scan to CAD, which produces 2D drawings from point cloud data but no BIM object intelligence. It is also distinct from point cloud processing, which refers to registration and cleaning of raw scan data, a step that precedes BIM modelling but does not produce a model.
The primary use cases sit within the existing building stock: measured surveys for refurbishment, as-built documentation where original drawings are missing or inaccurate, heritage building records, clash detection preparation, and handover of facility management data. New-build applications exist but are less common, typically limited to construction verification against design intent.

[edit] 2. How point cloud to BIM works

The workflow begins with 3D laser scanning on site. A time-of-flight or phase-based scanner — common instruments include the Leica RTC360, FARO Focus, and Trimble X7 — captures millions of points per second to produce a registered point cloud with a positional accuracy typically between ±1/8 in and ±3/8 in (±3 mm and ±10 mm), depending on scanner specification and site conditions.
Once the point cloud is registered and cleaned, it is imported into Revit (or another BIM authoring tool) as a reference dataset. The BIM modeller then traces and constructs parametric objects directly over the point cloud. This is not an automated process. Each element — structural member, wall face, pipe run, suspended ceiling — is modelled by a trained technician who interprets the scan geometry and applies BIM logic to it.
The modelling stage is where scope definition directly affects output quality. A model built to LOD 200 records general forms and approximate dimensions, adequate for space planning. LOD 300 models carry enough geometric precision for coordination and quantity take-off. LOD 350 adds interface geometry between systems. LOD 400 is construction-ready. LOD 500 designates an as-built record model with verified field conditions and non-geometric data, used for facilities management and asset lifecycle tracking.

[edit] 3. LOD specifications and deliverables

In the UK, LOD is defined within the ISO 19650 framework and the UK BIM Framework guidance. In the US, the BIM Forum LOD Specification (based on AIA G202) is the primary reference. Both frameworks describe LOD 100 through LOD 500, but the definitions are not identical between markets — practitioners working across jurisdictions should confirm which specification applies to each project.
A well-scoped point cloud to BIM commission will state, at minimum: the required LOD per discipline (architectural, structural, MEP), the coordinate system and survey control to be used, the file format for delivery (RVT, IFC, DWG, NWC), and whether COBie data output is required. In the UK, COBie is governed by BS 1192-4; in the US, the NBIMS-US V3 standard applies.
Typical deliverables from a point cloud to BIM service include the native Revit (.RVT) model file, an IFC export for open-standard interoperability, 2D drawing sheets extracted from the model, and a clash detection file in .NWC format for Navisworks review. Some projects also require a federated model combining architectural, structural, and MEP discipline models as separate linked files within a common Revit environment.

[edit] 4. Common applications by building type

Commercial office buildings. Refurbishment design, base-build verification, tenant fit-out coordination, and energy modelling inputs are the primary drivers. Floor-to-floor heights, structural grid, and core dimensions are the critical capture areas. LOD 300 is the standard minimum for design coordination.
Industrial facilities and manufacturing plants. Process plant documentation, pipe routing verification, equipment clearance modelling, and PDMS/BIM integration. MEP and structural content carries high density; clash detection against proposed new plant is a common deliverable. LOD 350–400 is typical.
Healthcare facilities and hospitals. Infection control requirements mean access to occupied areas is restricted. Phased scanning is common. BIM models feed directly into refurbishment planning and FM asset registers. COBie output is frequently required.
Heritage and listed buildings. Scan data captures irregular geometry that cannot be dimensioned manually with acceptable accuracy. Structural fabric, ornamental features, and existing services are modelled to client-defined tolerances rather than a fixed LOD.
Warehouses and logistics facilities. Floor flatness surveys, rack layout planning, and structural condition assessment. Models at LOD 200–300 are the norm for this building type.

[edit] 5. Data formats and software

The dominant authoring platform for point cloud to BIM is Autodesk Revit, which has native support for point cloud data via the RCP/RCS file format (Autodesk Recap). Other authoring environments — ArchiCAD, Bentley AECOsim — accept point cloud data but Revit holds the largest market share for this workflow, particularly in the UK, US, and Australia.
Point cloud registration and cleaning is handled in software such as Autodesk Recap Pro, Leica Cyclone Register 360, FARO Scene, or Trimble RealWorks, before the data is passed to the modelling team. These processing steps are typically handled by the scanning firm, not the BIM modelling provider.
Standard output formats are:

[edit] 6. Selecting a service provider

Point cloud to BIM is a modelling discipline, not a scanning discipline. The two are often conflated, but they require different skill sets. A laser scanning firm captures field data; a BIM modelling provider converts that data into a structured model. Many scanning firms outsource the modelling stage to specialist teams with Revit expertise.
Key factors to assess when selecting a modelling provider include: demonstrated experience at the required LOD across the relevant building types, familiarity with the applicable BIM standard (ISO 19650, BIM Forum LOD Spec, or project-specific BEP requirements), turnaround capacity relative to scan data volume, and QA/QC process for model accuracy checking against the source point cloud.
A pilot project — typically a single floor or a representative area — is the standard method for verifying provider capability before committing to a full-building scope. Model accuracy, element classification, parameter naming, and adherence to the agreed LOD are the checkpoints.

[edit] References / Further reading

• ISO 19650-1:2018 — Organisation and digitisation of information about buildings and civil engineering works, including building information modelling
• BIM Forum LOD Specification (current edition) — bimforum.org
• UK BIM Framework — ukbimframework.org
• BS 1192-4:2014 — Collaborative production of information, Part 4: Fulfilling employer's information exchange requirements using COBie
• Point cloud to BIM services — ViBIM, practical service overview covering Revit modelling from laser scan data for existing building and as-built projects