Bias and Repeatability of Measurements from 3D Scans Made Using iOS-Based Lidar

Abstract

<div class="section abstract"><div class="htmlview paragraph">Apple introduced a lidar-based depth sensor and enhanced Augmented Reality (AR) application programming interface (API) to the 2020 iPad Pro and iPhone 12 Pro, making widespread use of 3D scanning possible. Here we quantified the bias and repeatability of the 3D scans made using this system. The exterior and interior of a single vehicle and the exterior of a filing cabinet were scanned four times by eight different operators. Each operator then extracted four measurements from each of the exterior scans, five measurements from each of the interior scans, and three measurements from each of the filing cabinet scans. Hand measurements using a string and tape measure were used as reference values to estimate the bias. The values extracted from the 3D scans were biased between 0.9 and 9.3 cm below the actual exterior measurements and between 2.2 cm below and 0.3 cm above the actual interior measurements. The repeatability standard deviation varied from 1.8 to 5.8 cm for the exterior measurements and from 0.6 to 1.4 cm for the interior measurements. Based on the measurements we used here, the interior 3D scans were more accurate than the exterior scans, with the interior scans having a similar bias, though larger variance, to hand measurements.</div></div>