■ MEP BIM INSIGHTS — POINT CLOUD

How to Prepare Your Point Cloud for Revit: File Formats, Registration, and Common Errors

Getting a point cloud into Revit is not complicated. Getting it in correctly — with the right file format, accurate registration, proper coordinate alignment, and manageable file size — requires preparation that most teams skip when they are under schedule pressure.

The consequences of skipping that preparation show up later: a point cloud that loads but sits 200 feet from the building model, scan positions that drift between floors, file sizes that crash Revit every time a section is opened, or a model that comes back from the BIM team with everything at the wrong elevation.

This article covers the full preparation workflow — from raw scan data to a Revit-ready point cloud that a BIM team can model from reliably.

---

Step 1: Understand the File Format Options

Point cloud data comes off the scanner in proprietary formats specific to each manufacturer. Before anything else, that raw data needs to be processed into a format Revit can work with. There are three options:

Format	Description	Best for
.RCP	Autodesk ReCap Project file. Container that references multiple .RCS scan files. Native to Revit — links directly without conversion.	Standard workflow when all team members use Autodesk tools
.RCS	Autodesk ReCap Scan file. Individual scan region within an .RCP project. Can be linked to Revit directly but without the full project structure.	Linking a single scan region into Revit for a specific scope area
.E57	Open exchange format (ASTM E2807). Widely supported across scanning software, BIM platforms, and point cloud processors. Requires import into ReCap before linking to Revit.	When scan data comes from non-Autodesk scanners or software; open BIM workflows

---

Step 2: Registration — The Critical Step Most Teams Rush

Registration is the process of aligning multiple scan positions (each scan covers a limited area) into a single unified point cloud. A scanner placed in a corridor captures that corridor. A scanner in the adjacent room captures the room. Registration stitches them together so they read as one continuous dataset.

Poor registration is the most common source of point cloud errors that cause problems in the BIM model. When registration is off, walls are not straight, floors are not flat, and pipe runs that should be continuous appear broken or offset between scan positions.

Registration methods

Target-based registration uses physical targets (checkerboard panels or sphere targets) placed in overlapping areas between scan positions. The software matches targets across scans to align them. This is the most accurate method and is recommended for MEP retrofit projects where ±10mm accuracy is required.

Cloud-to-cloud registration matches the geometry of the scan data itself — the software finds matching features in overlapping areas between scan positions and aligns them automatically. Faster to set up but less accurate than target-based registration, particularly in open spaces with few geometric features.

Verifying registration accuracy

Every registration software produces an error report showing the residual misalignment between scan positions. For MEP BIM modeling, the target accuracy is:

• RMS error below 6mm between adjacent scan positions
• Maximum error below 10mm at any individual target
• Bundle adjustment error below 3mm across the full project

If the scanning contractor provides a point cloud without a registration report, ask for one before starting modeling. A registration error of 20mm or more will produce a model that cannot be used for coordination.

---

Step 3: Coordinate System Alignment

The point cloud must be aligned to the same coordinate system as the Revit project before it is sent to the BIM team. This is the step most frequently skipped — and the one that causes the most rework.

A point cloud in its scanner coordinate system (arbitrary origin, arbitrary north) will import into Revit at a location that has no relationship to the project base point or survey point. The result is a model and point cloud that exist in different places. Aligning them after the fact, with modeling already started against an incorrectly placed cloud, is significantly more work than doing it correctly before modeling begins.

How to align correctly

• Establish a known reference point in the building (a column centerline intersection, a survey control point, or a benchmark) and record its coordinates in the building coordinate system
• Identify that same point in the registered point cloud in ReCap
• Apply a transformation in ReCap to move the point cloud so the reference point aligns with its correct coordinates
• Verify by checking at least two additional known points after transformation — if only one point matches, the rotation is likely wrong
• Export the aligned .RCP and confirm the coordinate values before sending to the BIM team

---

Step 4: Cleaning and Organizing the Point Cloud

Raw point clouds contain data that is useful for the scanning team but not for the BIM modeler: scanner positions themselves (captured as dense spherical clusters), temporary objects (scaffolding, site equipment, people who walked through during scanning), and noise artifacts from reflective surfaces.

Before delivering to the BIM team, clean the cloud in ReCap or the scanning software:

• Remove scanner station spheres and target artifacts
• Remove temporary objects that are not part of the permanent building
• Clip the cloud to the project extents — exterior scans, parking areas, and roof scans not needed for the MEP scope should be excluded to reduce file size
• Organize by floor or zone using ReCap regions so the modeler can load only the relevant portion of the cloud in Revit

File size management

Large unoptimized point clouds cause Revit performance problems. A full-building scan for a 200,000 sq ft facility can exceed 20GB in raw format. After processing and optimization in ReCap, the same data should be under 5GB as an .RCP file.

If the processed .RCP is still very large, split it by floor as separate .RCS regions within the project. The modeler can link only the floors relevant to the current modeling scope.

---

Common Errors and How to Avoid Them

---

Checklist Before Sending to the BIM Team

• Registration completed with target-based method; error report generated and reviewed
• All scan positions show RMS error below 6mm
• Point cloud aligned to building coordinate system; verified at two or more reference points
• Temporary objects and scanner artifacts removed
• Cloud clipped to project extents and organized by floor in ReCap regions
• Exported as .RCP (not raw scanner format, not .E57)
• File size verified — full project under 5GB; individual floor regions under 1GB
• Test link into a blank Revit project to confirm correct placement before delivery

What We Receive at GEOMETRY-S

We accept point cloud data in .RCP, .RCS, and .E57 formats. For .E57 files, we handle the ReCap conversion and alignment verification as part of our project intake process at no additional charge.

If you are unsure whether your point cloud is ready for MEP BIM modeling, send us the file and we will review it and identify any preparation steps needed before modeling begins.