iPhone LiDAR for property inspectors

1. What the iPhone LiDAR sensor actually is

The LiDAR Scanner introduced on the iPhone 12 Pro and continued on every Pro-tier iPhone since is, technically, a direct time-of-flight VCSEL-array sensor that emits a structured pattern of infrared light pulses and measures how long each pulse takes to return to the sensor. Multiplying that round-trip time by the speed of light gives a distance estimate per illuminated point. The sensor produces a sparse depth map at a few hundred discrete points per frame, refreshed continuously, and ARKit fuses successive frames into a denser room-scale point cloud as the user walks the device through the space.

The important distinction from a survey-grade LiDAR scanner — the kind a quantity surveyor mounts on a tripod and pays five figures for — is that the iPhone's sensor was designed for augmented-reality placement, not for measurement. The marketing materials emphasise speed and integration with the camera, not millimetre accuracy, and the sensor's actual operating envelope reflects that priority. It is a phenomenally capable consumer sensor for the price; it is not a metrology instrument.

2. Sub-cm vs cm-level accuracy: the gap between the headline and the field

The headline number frequently quoted in iPhone LiDAR marketing and in third-party reviews is "sub-centimetre accuracy at typical room ranges." The headline is technically defensible for a single laser-spot reading on a cooperative diffuse surface at close range — under a metre, perpendicular incidence, white matte plaster, normal indoor temperature. The same sensor in a real property inspection regularly produces room dimensions that disagree with a tape measure by one to three centimetres at three to five metre ranges.

The reasons are mundane and worth understanding because they predict when LiDAR will fail you on a real inspection. The sensor's per-spot accuracy degrades non-linearly with range — at three metres, the sensor is competing with photon noise from ambient lighting, the dot pattern is sparser per unit of wall area, and any non-perpendicular incidence (a wall scanned at a shallow angle) reduces the reflected-photon count and inflates the noise on each measurement. ARKit's software stack is doing real-time mesh reconstruction with strong priors, which means the displayed room corners look right even when the underlying point cloud is noisier than the headline suggests. The buyer of an iPhone LiDAR scan is paying for the integration; the metrology is, frankly, what it is.

3. Where iPhone LiDAR fails on a real inspection

Three failure modes recur on every inspection long enough to encounter them. Large rooms: any single dimension over the LiDAR's effective range — Apple's published spec is up to 5 metres — produces increasingly noisy depth estimates beyond about three metres and unreliable estimates beyond five. A typical Portuguese sala or open-plan kitchen-living area routinely runs eight to twelve metres on the long axis, and the LiDAR-only scan of such a room will produce a mesh that visually looks fine but contains room-dimension errors of several centimetres. The fix on Inspecto's workflow is automatic anchor placement at corners using multi-frame fusion plus a laser-distance-meter cross-check for any room over five metres on the long axis.

Glass and mirrors: every reflective or transparent surface returns either no infrared echo or a misleading one. A plate-glass window facing a garden returns the wall behind the garden, not the plane of the glass. A mirrored bathroom returns a doubled room with the mirror plane absent from the mesh. The sensor hardware cannot detect these failure modes intrinsically, and ARKit's software stack has limited heuristics to compensate. The Inspecto workflow flags glass and mirror surfaces at scan time so the inspector can manually confirm them with a tape measure or a laser distance meter, and the room model treats the unconfirmed regions as "approximate" until they are verified.

Dark, glossy, or near-black surfaces: very low albedo to the sensor's near-infrared wavelength returns weak echoes that get drowned in ambient noise. Black slate floor tiles, dark hardwood flooring, glossy black appliances, and unlit cellars all produce noisy or missing depth in the point cloud. Bright daylight from a window can also overload the sensor at certain angles, producing range under-reads on surfaces in the same frame. The fix is the same as for glass: flag at scan time, cross-check with a tape, mark the region as approximate in the model.

4. Comparison to a laser distance meter

A modern handheld laser distance meter — the Bosch GLM, Leica Disto, or Stabila kits inspectors typically carry — measures a single-spot distance to a single target with quoted accuracy of about one to one and a half millimetres at typical room ranges. A single laser-distance reading is more accurate than any single iPhone LiDAR reading by a factor of around ten. The trade-off is throughput: a tape or laser meter measures one dimension at a time, while iPhone LiDAR captures hundreds of thousands of points across the whole room in seconds.

The right way to think about the two tools is therefore complementary rather than competing. iPhone LiDAR is the right tool for the visual model of the property — a buyer-shareable 3D walkthrough, a heat-map-anchored room mesh, a quickly produced floor plan accurate enough to specify a kitchen renovation. The laser distance meter is the right tool for any single load-bearing dimension that has to be defensible in a price negotiation: the ceiling height that determines whether a buyer can fit a specific staircase; the door width that determines whether the buyer's existing furniture passes through; the room dimension that drives the plot-area number on the deed. The Inspecto workflow uses LiDAR for everything by default and flags any single dimension over five metres, any room with notable glass or mirror surfaces, and any structural dimension that ends up in the buyer's offer letter, for laser cross-check.

5. Best practice during the scan

Six practical rules turn an average iPhone LiDAR scan into a defensible one. One: walk the room slowly. ARKit fuses frames over time, and a slow walk produces a denser, lower-noise mesh than a sweep at running speed. The published guideline is about one metre per second, but slower is fine and faster is not.

Two: keep the sensor pointed at where you want depth. The LiDAR's field of view is roughly the rear camera's field of view, and points outside that cone do not get measured no matter how long you stand there. Three: avoid scanning into bright daylight from windows. Shoot from the window outward, not toward the window, where the sensor's ambient-light rejection is weakest. Four: scan in good interior lighting. The sensor itself is active (it provides its own infrared illumination) but ARKit's feature-tracking visual side relies on visible-light features for frame-to-frame alignment, and a dark room produces drift even when the LiDAR depth itself is fine.

Five: cover every wall from at least two angles. Single-angle scanning leaves blind regions behind every furniture occlusion, and ARKit's mesh reconstruction will fill the gap with a guess rather than data. Walking the wall twice — once close, once from across the room — produces a noticeably better mesh than a single pass. Six: anchor the model with a laser-distance reading on at least one major dimension. The Inspecto workflow takes a confirmed laser reading at the diagonal of every scanned room and uses it to scale-correct the LiDAR mesh; this single step turns a 1–3 cm RMS scan into a sub-1 cm scan for the corrected dimensions, at almost no time cost.

6. iPhone LiDAR vs survey-grade scanners

A professional survey-grade scanner — a Leica BLK360, a Faro Focus, a NavVis VLX — produces millions of points per second, achieves quoted accuracy of a few millimetres at ten-metre ranges, and produces a point cloud dense enough to drive structural analysis or full BIM-ready model generation. They cost ten to thirty thousand euros, weigh several kilograms, and require professional training to deploy productively. They are the right tool for a renovation contract that has to specify millimetre tolerances on a structural alteration; they are over-specified for a residential pre-purchase inspection where the buyer needs a visual model and a defensible floor plan.

The iPhone LiDAR's real value to property inspection is the price-performance gap. For most residential pre-purchase work, the buyer cares about the visual room model, the major dimensions, and the heat-map-anchored damp evidence — none of which require a millimetre of accuracy on every wall. The iPhone's 1–3 cm RMS, with a laser-distance scale anchor at the diagonal, is genuinely sufficient for that workload, and the integration with the camera, the mesh-export pipeline, and the moisture and photo overlay tools is what makes the workflow practical. Where the buyer needs survey-grade precision — a structural alteration, a planning submission requiring an architectural drawing, a deed-defining boundary — the inspection report flags the dimensions that need it, and the buyer commissions the appropriate specialist as a follow-up.

7. What the inspection report turns the scan into

A finished Inspecto inspection turns the iPhone LiDAR scan into four artefacts the buyer keeps for life. The first is a 3D walkthrough mesh, viewable in any modern browser, that lets the buyer revisit any room of the property weeks after the inspection visit — useful for sharing with a contractor, an architect, a family member, or any party the buyer needs to consult before signing the deed. The second is a 2D floor plan extracted from the mesh, scaled against the laser-distance anchor, with room labels and the major openings marked. The third is the moisture heat map anchored to the mesh, so a wet patch on a wall in the report ties to a specific cell of a specific wall of a specific room rather than to a generic photo. The fourth is the structural and 84-point findings tied spatially to the mesh — the chimney flashing photographed from outside ties to the chimney visible in the loft scan, and the rising-damp band on wall N3 ties to the rising-damp band on the heat map.

That spatial integration is what makes the scan worth doing rather than a sequence of disconnected photos. A buyer with a 3D model, a measured floor plan, a heat map, and an 84-point checklist all anchored to the same room mesh has a single source of truth for the property; a buyer with a folder of photos and a ten-page Word document has a much weaker hand at renegotiation. For the broader workflow that produces these four artefacts from the raw scan, see the LiDAR scanning workflow; for where the scan fits into the pre-purchase timeline, see the pre-purchase inspection guide.