- 01The workflow is: capture → register & clean the point cloud → set shared coordinates → model by discipline to the target LOD → QA against the cloud → deliver.
- 02Registration quality (aligning scans into one accurate cloud) sets the ceiling on model accuracy — errors here propagate everywhere.
- 03Shared coordinates and levels must be established before modeling so the model aligns with other project data.
- 04QA is a deviation check of the model against the cloud, not a visual glance — this is where accuracy claims are proven.
- 05AI classification pre-segments the cloud and pre-places common elements, cutting manual modeling time.
Turning a laser-scanned point cloud into a usable Revit model is a disciplined pipeline, and each step constrains the next. Skipping or rushing registration, coordinates, or QA is where Scan-to-BIM projects quietly go wrong — the model looks fine but doesn’t line up, isn’t accurate, or can’t be used downstream. Here’s the workflow done properly.
The end-to-end pipeline
- 01Capture
Terrestrial laser scanners (and increasingly mobile/handheld or drone LiDAR) capture the space as millions of 3D points, with overlap between scan positions for registration.
- 02Registration & cleaning
Individual scans are aligned into one coherent cloud (registration), then noise, people, and transient objects are removed. Registration accuracy caps the whole project’s accuracy.
- 03Import & shared coordinates
The cleaned cloud is brought into Revit (via ReCap) and the project’s shared coordinate system and levels are established so the model aligns with survey and other models.
- 04Modeling by discipline
Elements are modeled to the agreed LOD — architectural, then structural, then MEP — snapping to the cloud and using project families and standards.
- 05QA against the cloud
The model is checked for deviation from the point cloud within tolerance, and for standards compliance — the step that substantiates the accuracy you promised.
- 06Delivery
Native Revit, IFC, and any 2D extractions are issued with documentation of LOD, tolerance, and coordinate basis.
Accurate models, aligned data
A Scan-to-BIM model is only useful if it’s accurate and it aligns with everything else on the project. Spetia Engineering runs the full pipeline to standard — rigorous registration, correct coordinates, disciplined modeling, and real deviation QA — with AI absorbing the repetitive work so quality and speed both go up.
