How Accurate Is an iPhone LiDAR Floor Plan? (And How to Trust It)

In good conditions, iPhone LiDAR floor plans are usually accurate to within a few centimetres — good enough for real-estate floor areas, as-built drawings, space planning and maintenance scoping. But iPhone LiDAR accuracy is not a single fixed number, and treating it like one is how people get burned: the real figure shifts with the device, the room size, the surfaces and how carefully the space was captured. Before you rely on a dimension for anything contractual, structural or safety-critical, validate it against a tape or laser. The honest, more useful answer is that accuracy depends on the capture, and the tools that deserve your trust are the ones that tell you how accurate a given scan actually is — rather than handing you a figure and hoping you believe it. This article breaks down where the accuracy comes from, what changes it, and how calibration turns a good-enough number into one you can defend.

What "accurate to a few centimetres" really means

A LiDAR-equipped iPhone or iPad Pro builds the geometry of a space from millions of depth samples as you walk through it. Under good conditions the resulting measurements often land within a few centimetres of the truth — sometimes better, sometimes worse, depending on the device, the room and the capture. Any single headline figure actually hides two very different kinds of error, and understanding the split is the key to understanding iPhone LiDAR accuracy:

• Bias — a consistent, systematic offset, for example every length reading a steady one percent long. Bias is the good kind of error, because if you can measure it you can correct it away.
• Noise — random scatter from run to run, the physical noise floor of the sensor. Noise cannot be corrected out; it is the hard limit on how precise the instrument can be.

This matters because a tool that quietly corrects bias but hides noise is flattering itself. The trustworthy approach is to report both: how far off the scan tends to be, and how much irreducible scatter remains underneath. When you know the bias, you can subtract it; when you know the noise floor, you know how much run-to-run variation to expect and can decide whether the scan is precise enough for the decision in front of you. A number without that context is just a hopeful guess dressed up as a measurement.

What affects the accuracy of an iPhone LiDAR scan

Two identical phones can produce noticeably different results depending on how, and where, they were used. These are the factors that move the needle most:

• Coverage: walking the full perimeter and catching every corner beats a rushed pass that leaves the geometry to be guessed.
• Distance and room size: very large rooms and long unbroken spans accumulate more drift than compact, well-defined spaces.
• Surfaces and lighting: glass, mirrors and very dark or shiny surfaces are genuinely hard for LiDAR, and tagging them helps the software handle them correctly.
• Device and steadiness: a steady, chest-height walk at walking pace gives the sensor cleaner geometry than a fast sweep.

Don't just trust the number — measure it

Most measuring tools give you a figure and stop there. The problem is that a floor area you cannot defend is a liability: on a listing it can become a mis-description complaint, and in facilities it becomes a change order when a contractor's real measurements disagree with yours. The better approach is to characterise a scan's accuracy against known references — separate the bias from the noise floor, and check the result out-of-sample so the tool cannot grade its own homework. Then the number you act on comes with evidence behind it, not just a promise.

How calibration makes an iPhone LiDAR scan trustworthy

This is exactly what FIMIT's Calibrate-a-Scan does. You enter a few real-world reference measurements — lengths you have physically verified with a tape or laser — and FIMIT fits a size-weighted correction factor to your scan, complete with a confidence score and automatic flagging of any reference that looks like a mistake. That factor is applied to the measurements and exports you work from, correcting the systematic bias while staying honest about the noise floor that remains. The result is iPhone LiDAR accuracy you can quantify, not just claim.

Reversible by design

Calibration in FIMIT is additive and fully reversible — the underlying geometry is never altered, so a scan can be reverted to raw at any time, and every calibrated export is clearly labelled as calibrated. You get measurements you can stand behind, an honest read on how accurate they are, and nothing to unpick if you change your mind. Walk a space with an iPhone, calibrate it against a couple of known lengths, and publish a floor area with evidence behind it.

iPhone LiDAR vs a professional measured survey

A traditional measured survey, done by a surveyor with total-station or high-end laser gear, can reach millimetre-level precision on individual measurements — but it is slow, costly and hard to repeat across a whole portfolio. iPhone LiDAR trades precision for large gains in speed, cost and repeatability: anyone on the team can capture a space in minutes, as often as the building changes. The right question is rarely which is more precise in absolute terms; it is which gives you an accurate-enough, defensible number at a cost and cadence you can sustain. For listing documentation, everyday as-built records, space planning and quotation-stage measurement, the phone is often the practical choice — provided the capture is reviewed and critical dimensions are validated. Formal, contractual or higher-precision work still belongs with a professional measured survey.

Is it accurate enough for your job?

For much of everyday property and facility work, yes. Accuracy within a few centimetres sits inside the tolerance that typically matters for a listing's floor area, an operational as-built record, a cleaning or maintenance quote, or space planning — but suitability always depends on the specification, the building and the confidence the decision requires, so review the capture and validate the dimensions that carry risk. Where millimetre-level engineering precision is required — structural fabrication, for instance — no phone is the right tool, and where a contract, regulation or dispute demands a certified measured survey, commission one. The useful question is not whether LiDAR is perfect but whether it is accurate enough for this decision, and whether you can prove it. With characterisation and calibration, you can answer both.