A Practical Guide to Reality Capture Workflows

A reality-capture project rarely fails because the scanner, camera or drone was incapable of collecting data. More often, it fails because the team captured the wrong area, missed control, used inconsistent settings or handed over files the client cannot use. This guide to reality capture workflows sets out a practical route from defining the requirement to delivering reliable outputs, whether you are surveying an existing building, documenting a construction site or inspecting difficult-to-access assets.

Start with the required deliverable

The workflow begins before equipment leaves the office. Ask what decision the captured data must support. A point cloud for clash detection, a measured CAD drawing, a BIM model, a virtual property tour and a condition-inspection record all require different levels of accuracy, coverage and processing.

Agree the coordinate system, datum and units at the outset. On a larger site, establish whether the client needs local grid coordinates, national grid coordinates or a model aligned to an existing design file. If the capture is intended for repeat surveys, such as progress monitoring or asset management, consistent control is essential. A visually impressive model that cannot be compared with later data has limited operational value.

Define acceptance criteria in plain terms. This may include the required point spacing, positional tolerance, image resolution, model format, naming convention and delivery date. It is also worth agreeing what is outside scope. For example, a terrestrial laser scan may capture the visible face of a façade exceptionally well, but it will not record concealed structure behind cladding.

Choose a capture method that suits the site

Reality capture is not one technology. It is a combination of methods selected for the job, the environment and the level of confidence required.

Terrestrial laser scanning is well suited to detailed as-built surveys, complex plant rooms, heritage features and interiors where accurate geometry matters. It produces dense point clouds quickly, but requires careful scan placement to reduce shadows and occlusions. A scanner positioned in the middle of a room will not adequately record the back of pipework, recesses or areas behind furniture.

GNSS and total stations provide the control that gives wider projects a dependable spatial framework. They are particularly valuable where scan data, drone imagery and design models need to align. In obstructed urban locations, under tree cover or around tall structures, a total station may be the more dependable route to establishing control.

Drone photogrammetry can cover roofs, façades, stockpiles and large sites efficiently. It is productive, but weather, airspace restrictions, safe launch areas and adequate ground control all need consideration. Photogrammetry is also less forgiving of repetitive surfaces, reflective materials and poor light than many teams expect.

Handheld scanners and 360-degree cameras have a place where speed and accessibility are the priority. They can support facilities records, property marketing, preliminary surveys and rapid site communication. They should not automatically be treated as a replacement for survey-grade scanning where dimensional accuracy is contractually critical.

Plan the reality capture workflow before arriving on site

A short pre-start plan can prevent hours of remedial work. Review drawings, site photos and access information to identify likely blind spots, restricted zones, traffic routes and areas where reflective or moving surfaces may affect data quality. Consider when the site will be least busy. Capturing a live warehouse during peak movement can create incomplete scans and unnecessary safety exposure.

Build a simple capture plan that identifies control locations, scanner or camera positions, anticipated overlap and the order of work. In multi-storey buildings, plan how floors will be connected. Stairwells, lift lobbies and external control points are often useful links, but they must be captured clearly and consistently.

Site safety remains part of the technical process. A scan position may provide excellent coverage but be unsuitable if it blocks an escape route or sits within a plant movement zone. Follow the site induction, use appropriate exclusion areas and consider a spotter where operating close to vehicles, lifting operations or the public. For drone operations, assess the operational area, people on site and local restrictions before deployment.

Check batteries, storage media, targets, tripods, chargers and field-control equipment before travel. It sounds basic, but a missing tribrach or full memory card can compromise a planned survey day. Where project risk is high, carry spare power and a second method of recording essential control observations.

Establish and verify control

Control is the backbone of a survey-grade workflow. Set out clearly visible, stable points that will not be disturbed during capture. Record enough redundancy to check the work rather than simply making it fit. If one control point appears inconsistent, a network with independent observations gives you a basis for investigation.

Use targets or natural features only where they can be identified unambiguously in the data. Targets should be distributed through the project, not clustered around the first scan position. For drone photogrammetry, ground control points need to be visible in imagery and measured accurately. Check points that are excluded from processing provide a more honest indication of final accuracy.

Capture systematically, not just quickly

Work to a repeatable sequence. At each station or flight stage, confirm that planned areas are covered, that overlap is sufficient and that the captured data is readable. Do not wait until back at the office to discover that a key elevation was obscured by a parked lorry or that a scanner was set to an unsuitable resolution.

Scanner settings should reflect the required output. Higher resolution and quality settings can improve detail, but they increase field time, file size and processing demands. For a broad building shell, excessively dense data may add cost without improving the final drawing. For intricate steel connections, ornate heritage fabric or plant interfaces, the extra detail may be justified.

Record field notes as you go. Note access limitations, areas not captured, temporary obstructions, scan names, control IDs and any changes to the plan. Photographs of control points and challenging locations are useful when a processor needs to understand an unexpected gap or alignment issue days later.

Carry out a field check before demobilising. Review coverage on the device where possible, inspect control observations and confirm that critical features have been captured from more than one position. A twenty-minute check on site is usually cheaper than arranging a return visit after trades have progressed or access has been removed.

Process data with quality checks built in

Processing should follow a controlled sequence: transfer and back up raw files, register scans or align images, apply control, clean obvious noise, assess accuracy, then create the agreed deliverables. Keep raw data separate from working and issued files. A clear folder structure and version naming convention reduce the risk of overwriting source information or issuing an outdated model.

Registration reports should be reviewed rather than accepted at face value. Low residual values are encouraging, but they do not prove complete coverage or correct georeferencing. Check known dimensions, inspect overlaps, compare independent check points and look closely at difficult areas such as glass, shiny metal, narrow corridors and long featureless walls.

Cleaning requires judgement. Remove obvious moving objects, isolated points and unwanted background where it improves usability, but avoid over-cleaning. Deleting points around edges, services or structural interfaces can remove information that designers and engineers need. Retain an original registered point cloud before producing a simplified version for sharing.

The right deliverable depends on the end user's software and capability. A full-resolution point cloud may be ideal for a survey or design team, while a decimated cloud, orthomosaic, mesh, panoramic tour or measured drawing may be more practical for a facilities manager. Confirm file formats early, particularly where the data will feed into BIM coordination, machine control or a common data environment.

Deliver data people can act on

A well-organised handover makes captured information useful beyond the immediate project. Issue the files with a brief survey report that states the coordinate system, control method, estimated accuracy, capture dates, exclusions and any known limitations. Include a sensible file structure so the recipient can identify raw data, registered point clouds, derived drawings and final models without guesswork.

For larger datasets, consider how the client will view and share the information. Not every stakeholder needs to download a dense point cloud. A lightweight viewer-ready format or clearly named clipped areas can make the data more accessible while keeping the authoritative source intact.

Reality capture also needs a retention plan. Construction teams may need the data for claims, progress evidence or verification. Building operators may return to it years later when planning refurbishments. Retaining the original data, processing settings and control records protects that future value.

Build a workflow your team can repeat

The most effective workflows are not necessarily the most complicated. They give field teams a clear plan, give processors traceable information and give clients data that fits their next decision. Standard operating procedures, equipment familiarisation and occasional refresher training help maintain that consistency as projects and staff change.

There is no single scanner, drone or camera that is right for every project. The practical decision is whether the equipment, accuracy, support and commercial model fit the work in front of you. For teams testing a new method or scaling a busy programme, Survey Tech can help assess suitable capture equipment, hire options, training and ongoing servicing - so the workflow remains dependable long after the first survey day.