How accurate are AR-generated floorplans?

On LiDAR-enabled devices, SiteSurvey AR floorplans typically achieve dimensional accuracy within 1 to 2% of the true measurement for rooms up to 20 metres in length. For a 10-metre wall, this translates to a variance of roughly 10 to 20 centimetres, which is acceptable for the majority of commercial survey, space planning, and fit-out applications.

Accuracy is influenced by the device used, the quality of the capture technique, and environmental conditions. LiDAR-equipped devices (such as the iPad Pro and iPhone Pro range from 2020 onwards) produce more reliable results than devices relying on computer vision alone. Good lighting and unobstructed surfaces improve consistency, while glass, mirrors, and very dark or reflective walls can introduce additional variance.

For surveys where tighter tolerances are required, SiteSurvey supports the addition of manual dimension overrides. A surveyor can measure a key wall with a laser measure and enter the corrected value into the plan, which the app then uses to recalibrate surrounding geometry. This hybrid approach combines the speed of AR capture with the precision of targeted manual measurement.

It is worth noting that AR floorplan accuracy is comparable to, and in many cases better than, the typical variance introduced by manual tape measurement over long distances or around obstacles. For schematic surveys, space utilisation assessments, and condition reporting purposes, AR-generated plans provide a reliable and consistent spatial record.