Why the point cloud format decides planning usefulness
Photogrammetry from supplied imagery can produce textured meshes, orthophotos, point clouds, CAD traces, BIM-oriented models and viewers. The point cloud is not the final value by itself. It is a reference or intermediate dataset for real planning work.
Autodesk describes RCP as a project file that references RCS point cloud data. E57, PTS and PCG are described in ReCap as single scan files that can package project information. Pix4D lists LAS, LAZ, PLY and XYZ among point cloud outputs. The practical conclusion is simple: the right delivery depends on the receiving software.
That matters for Voxelia because our work is not selling drone flights. We review existing imagery and turn it into usable planning data. A clean delivery must state whether the dataset is for Revit reference, CAD measurement, GIS positioning, scan-to-BIM, PV roof modeling, a viewer or quality control.
E57, LAS/LAZ, RCP/RCS, PLY and XYZ compared
ASTM E2807 describes E57 as an exchange format for 3D imaging data. It can store 3D points, attributes such as color and intensity, and 2D imagery from a 3D imaging system. That makes it useful when data must move between software ecosystems.
LAS and LAZ are common for large, georeferenced point clouds in surveying and GIS. LAZ is compressed and therefore practical for transfer and archiving. For buildings, facades and roof models, units, origin, coordinate system and point density matter as much as file size.
RCP and RCS are common Autodesk target formats. Revit and AutoCAD workflows often reference ReCap-indexed point clouds. PLY and XYZ are lighter exchange formats, but depending on export settings they can lose structure, coordinate reference or attributes.
| System / dataset | Suitability | Best for | Practical note |
|---|---|---|---|
| E57 | Neutral point cloud exchange | Scan-to-BIM, quality review, cross-software delivery | Useful when the receiver is not only using Autodesk or needs an archive-friendly exchange dataset. |
| LAS/LAZ | Georeferenced point clouds and large datasets | GIS, terrain, orthophoto context, surveying deliveries | Strong for spatial context and large datasets; usually not the direct final state for Revit. |
| RCP/RCS | Autodesk delivery | Revit, AutoCAD and Navisworks-oriented coordination | Often the right target when the point cloud is meant to be referenced directly in Autodesk software. |
| PLY | Colored point cloud or 3D object-adjacent workflows | Viewers, research, mesh-adjacent workflows, Gaussian-splatting-adjacent checks | Check whether scale, coordinates and attributes are enough for planning. |
| XYZ / TXT | Simple raw data exchange | Quick checks, scripts, special imports | Only useful with clear column order, units, coordinate system and attribute notes. |
Which formats fit Revit, AutoCAD, Archicad, QGIS and viewers
Revit and AutoCAD workflows often require a ReCap step because point clouds are referenced as RCP/RCS projects. Autodesk describes RCP as a project file that references RCS scan files, so delivery is often a folder package, not just one file.
GIS workflows often prefer LAS/LAZ point clouds and GeoTIFF orthophotos over an Autodesk-only package. Coordinate system, vertical datum, units, classification and tiling matter more than a photorealistic viewer impression.
Archicad and openBIM workflows depend on import route, project origin and intended downstream use. Often the point cloud is not the final product; the end value is a BIM model, an IFC delivery, facade plan, roof model or CAD trace derived from it.
Preserving coordinates, color, intensity, units and point density
A photogrammetry point cloud is more than XYZ coordinates. Planning and acceptance depend on color, point density, coordinate reference, vertical datum, units, local origin, transformation notes and documented limits.
The ASTM summary states that E57 can store 3D point data, attributes including color and intensity, and 2D imagery. In photogrammetry, intensity is not automatically comparable to laser scanner intensity; RGB color from imagery is often more relevant. LAS/LAZ can be strong for georeferenced clouds but still needs correct scale, tiling and documentation.
Voxelia therefore documents the output purpose: local building model, georeferenced orthophoto, CAD reference, Revit point cloud, PV roof model or web viewer. That prevents a dataset from being used for decisions it was not checked for.
Common point cloud delivery mistakes
Many problems happen during delivery rather than reconstruction. A point cloud can be correctly processed and still arrive in CAD with the wrong scale, origin or attributes.
| Risk scenario | Why it matters | Typical symptom | Countermeasure |
|---|---|---|---|
| RCP without support folder | The project file references point cloud data and may not contain the points itself | Revit or AutoCAD shows missing or empty references | Deliver RCP/RCS as a complete package and preserve folder structure |
| Missing coordinate system | The receiver cannot verify location, elevation or project origin | The point cloud appears far from the model or rotated incorrectly | Document CRS, vertical datum, origin and transformation notes |
| Wrong point density | Very large clouds can be hard to work with without improving planning value | Slow import, poor navigation, messy CAD tracing | Thin, tile or convert the dataset according to the downstream task |
| XYZ without schema | Units, column order and attributes are ambiguous | Import is distorted or color is missing | Provide schema, units, decimal convention and attribute columns |
How Voxelia prepares point clouds from supplied imagery
The workflow starts with the intended output, not with a drone flight. We determine whether a point cloud is the main deliverable or whether CAD, orthophoto, mesh, BIM or viewer data will create more value.
Clarify target workflow and receiver software
Revit, AutoCAD, Archicad, QGIS, CloudCompare, PV tools and viewers all need different deliveries.
Review imagery, scale and coordinates
Sharpness, overlap, EXIF/XMP, reference measurements, GCPs and scale determine the reliability of the output.
Generate the right point cloud output
Depending on the workflow, the delivery can include E57, LAS/LAZ, PLY, XYZ or Autodesk-oriented RCP/RCS.
Adapt point density and data size
CAD tracing, scan-to-BIM, facades, roofs and viewers need different density, clipping and tiling choices.
Deliver notes and limits
Format, coordinates, units, quality limits, import notes and recommended downstream products are documented.
What a good delivery package should include
A good point cloud package contains more than a large file. It should include a README, delivery format, optional neutral exchange format, coordinate and unit notes, preview, quality notes and clear usage limits.
For Autodesk workflows, the receiver should know which RCP file to open and which RCS files belong to it. For E57 or LAS/LAZ deliveries, the receiver should know whether the cloud is local or georeferenced, which elevation reference was used and whether color attributes are present.
The short practical summary is: E57 is often the neutral exchange anchor, LAS/LAZ is strong for georeferenced point clouds, and RCP/RCS is often the Autodesk target state. The decisive factor is always suitability for the intended use.