What is a point cloud survey

A point cloud survey is a 3D record of an object, building, or site captured as millions of measured points in space. Each point has X, Y, and Z coordinates and usually intensity and colour information. The full set of points makes up a measurable, photorealistic record of the captured space. Point clouds have become the standard deliverable for as-built surveys, façade capture, structural monitoring, and BIM-ready geometry capture in UK construction.

What it actually contains

A typical point cloud from a building survey contains millions to billions of points. Each point is a measurement of where the laser pulse (LiDAR) or photogrammetric reconstruction located a piece of surface. Together, the points represent every visible surface in the captured space.

Beyond the position, each point carries:

• Intensity — how strongly the surface reflected the laser pulse (LiDAR) or how the surface appeared in the imagery (photogrammetry). Intensity reveals materials, edges, and texture.
• Colour (RGB) — the true colour of the surface, sampled from a camera. Colour makes the point cloud readable as if it were photographs in 3D.
• Classification — a label that identifies which kind of surface the point belongs to (ground, vegetation, structure, etc.). Classification is added in processing and makes downstream filtering possible.

A registered point cloud has all of these properties tied to a coherent coordinate system across many capture stations. The result is a single dataset that is geometrically accurate to a few millimetres and visually readable at any zoom level.

How it is captured

Two main methods produce point clouds:

Terrestrial LiDAR. A scanner on a tripod captures from each station for a few minutes. Multiple stations are required for any reasonably sized space — small spaces might need 5–10, a large building 100 or more. Stations are tied together using control targets or natural features.

Drone photogrammetry. A drone flies a planned grid over the site, taking many overlapping photos. Software processes the photos into a dense point cloud, typically registered to ground control points placed before the flight.

Specialist methods include mobile LiDAR (a vehicle-mounted scanner walking through a space), handheld scanners for tight or moving environments, and combined LiDAR-and-photogrammetry payloads.

Deliverable formats

A point cloud is delivered in industry-standard formats. The common ones:

• E57 — open, vendor-neutral, supported by every major design platform.
• RCP / RCS — Autodesk’s format, directly importable into Revit and other Autodesk products.
• PLY, PTS, LAS — used in specialist downstream pipelines.

Most projects receive an E57 or RCP/RCS as the primary deliverable, with optional drawings (floor plans, sections, elevations) or BIM models derived from the cloud.

What you do with it

A point cloud is a measurable, photorealistic record of the as-built. The downstream uses are wide:

Design. The cloud is the existing-conditions basis for any design work. Architects, engineers, and MEP designers all import the cloud into their respective platforms and design against it.

Clash detection. In a federated BIM environment, the point cloud is the existing-conditions reality the design must coordinate with.

Verification. Installed work can be verified against the design by overlaying the relevant model on the cloud.

Floor plans, sections, elevations. Drawings can be generated directly from the cloud at any time, including drawings that nobody anticipated when the survey was commissioned.

Modelling. A Revit or other BIM model can be built from the cloud at agreed Levels of Development, on demand.

Inspection. Defects, geometry queries, and visual checks can be made on the cloud without going back to site.

Heritage record. For listed buildings, the cloud is part of the building’s archival record.

Monitoring. Repeat captures, tied to the same control, detect movement at engineering tolerance.

How to specify a good point cloud survey

A good brief for a point cloud survey covers:

• The area to be captured.
• The accuracy required (a few millimetres for engineering work, a few centimetres for visual records).
• The deliverable formats required.
• Any specific conditions (security restrictions, working hours, occupied environments).
• Any downstream uses that will drive the brief (BIM modelling, listed-building documentation, structural monitoring).

A defensible deliverable includes:

• The registered point cloud in the agreed format.
• A registration report showing the cloud-to-cloud and cloud-to-control accuracies.
• Photographs of every scan station / flight position.
• A brief method statement.
• The names and qualifications of the surveyors.

If any of these are missing, ask. None are optional on engineering work.

Practical advice

Point cloud surveys are one of the most flexible deliverables in modern surveying. Once captured, the same cloud serves design, verification, dispute, and disposal — a one-shot capture with a long downstream life. The cost of a typical point cloud survey is a small fraction of the cost of any single design rework or remeasurement that the cloud will prevent.

For any project working with existing fabric, a point cloud survey early in the design phase is almost always money well spent.