3D laser scanning in Singapore.

What you receive

A registered, quality-checked point cloud of the building or site as it stands, in the format your tools read: E57 as the vendor-neutral default, LAS or LAZ for survey and GIS workflows, RCP/RCS for Revit and AutoCAD. Optional extras: mesh (OBJ/FBX/glTF), colourised output, and 3DGS scenes from compatible captures. Whole-building captures run to tens of gigabytes: we agree the delivery arrangement up front rather than emailing you a link that fails.

Alongside the cloud you get the things that make it usable six months later: a scan coverage note saying what was captured and what was not, the registration report, the coordinate system and units, and, where control was set, the control record. A point cloud without that context is a liability the next consultant has to re-do.

Handheld SLAM or tripod scanning

The honest trade-off: handheld SLAM wins on speed and access, tripod scanning wins on accuracy per station. An occupied office floor cannot be closed for tripod stations: the Lixel K2 captures it in a walk-through at 200,000 pts/s. Statutory or dispute-grade measurement wants the BLK360 G2’s 4 mm @10 m. Long corridors and external sightlines beyond 100 m call for the Lixel L2 Pro, which holds ≤1 cm relative within 100 m and reaches 120 m as standard. Most real projects combine methods, and combining them on one job is normal rather than a compromise.

A typical mixed job looks like this: handheld capture for the circulation, tenanted floors and back-of-house, tripod stations on the elements a drawing will be dimensioned from, grid lines, structural faces, lift cores, a facade bay, and the two datasets registered together against the same control. You pay for tripod time only where tripod accuracy earns it.

Accuracy, stated properly

Relative and absolute accuracy are different things, and conflating them is the most common way scan data disappoints. Relative accuracy is how correct a measurement is within the scan: the distance between two walls. Absolute accuracy is how correctly the whole scan sits in real-world coordinates. A device can be 1 cm relative and 3 cm absolute, and for many jobs only the first number matters. If your deliverable has to tie into a site grid, a Singapore coordinate system or a neighbouring survey, say so in the brief, that is an absolute-accuracy job and it changes how we set up.

Tripod accuracy is per station. The figure you actually get across a building depends on how well stations are registered together, so we quote the instrument figure with its condition attached and report the achieved registration afterwards. We do not publish a single flattering number and hope nobody checks it.

What we scan

• Whole buildings and single floors: offices, retail, industrial, institutional, residential.
• Occupied and operating spaces: capture around staff, stock and visitors, out of hours where needed.
• Plant rooms, risers and ceiling voids: the spaces nobody has drawings for and everybody needs measured.
• Facades and external envelopes, from the ground with long-range scanning, or from the air by drone survey where the elevation cannot be seen.
• Sites, yards and civil works: extents, levels and volumes, usually combined with topographic survey.
• Heritage and shophouse interiors: geometry that is irregular by nature and impossible to hand-measure honestly.

Site requirements

• Access to every space in scope: locked rooms not opened on the day are gaps in the data, and the second visit costs more than the key.
• Lighting is not required. LiDAR measures with its own laser; only colourised output benefits from light. Clients assume otherwise often enough that it is worth saying plainly.
• Occupancy is fine. Moving people, glass, mirrors, reflective surfaces and heavy clutter slow capture and add cleanup, but do not prevent it.
• Where absolute accuracy matters, we set control before capture and register the cloud to it.
• A site contact who can open doors, and a note of any permit, induction, hot-work or after-hours requirement: this is usually what determines the date, not our diary.

What slows a scan down

Worth knowing before you book, because it is what moves a half-day job to a full day. Mirrored and glazed surfaces return misleading points and need either avoidance or cleanup. Dense storage, stacked furniture and cable-tray congestion mean the scanner cannot see what is behind them, so occupied storerooms give partial data by definition. Long featureless corridors are hard for SLAM to hold onto; we plan loop closures through them rather than walking them once. Live traffic, forklifts and public circulation add passes. And splitting a building across several tenancies with separate approvals adds coordination time that has nothing to do with scanning.

Registration and QA before anything is delivered

Raw scans are not a deliverable. Every job goes through registration: aligning individual scans or SLAM trajectories into one coordinate system: followed by a check pass: cloud-to-cloud residuals, loop-closure checks on handheld data, control residuals where control exists, and a coverage review against the scope list. We remove obvious noise, stray people and vehicle passes, and confirm the file opens in the target software before it leaves. If a spec was not met, you hear it from us with the numbers, not from your modeller three weeks later.

What the point cloud feeds

Capture once, answer repeatedly. The same registered cloud is the source dataset for as-built drawings, for a Scan-to-BIM model at an agreed LOD, for a digital twin handover package, and for clash and fit checks your team runs directly in the cloud. Questions nobody thought to ask on the day are already answered in the data, so a second mobilisation is rarely necessary. It is also why we archive the registered dataset: the cheapest survey is the one you already own.

Turnaround and cost drivers

Capture is typically hours on site; processed deliverables for a normal floor plate arrive within a week. What moves the price: floor area and site extent, the number of separate spaces and tenancies, access and out-of-hours constraints, the accuracy specification (control and tripod work cost more than a walk-through), and the deliverables list. Repeat capture on a schedule, monthly progress records, for instance, prices differently from a one-off because the setup is already done.

Send the brief to get a quote: the address or drawing extent, the accuracy you need, the formats your team works in, and the date. We respond within one business day.

When scanning is not the answer

Three situations where we will tell you to spend the money elsewhere. First, when the official drawing already exists and is current: buying it is cheaper than measuring, and the BCA plan purchase guide explains the route. Second, when the question is about a single dimension in a single room: a tape measure and a photograph answer that, and paying for mobilisation, registration and QA to confirm one opening width is not good value. Third, when the real requirement is a legal boundary, that is a Registered Surveyor’s statutory work, and no amount of point-cloud density changes it: see land survey for where the line sits.

Conversely, scanning earns its cost quickly in four cases: the building has been altered and no drawing matches it; several deliverables will come from one visit, so the capture is amortised; the space is complex, irregular or occupied enough that hand measurement would be slow and error-prone; or the record needs to be defensible later, because a registered cloud can be re-measured by anyone, years afterwards, without a site visit.

We would rather tell you which category you are in during the first conversation than sell a capture that does not pay for itself. Send the question rather than the specification, and we will say which route is cheapest.