Indoor positioning systems (IPS) has seen an increase in demand because of the needto locate users in environments where Global Navigation Satellite Systems (GNSS) lacksaccuracy. The current way of implementing an IPS is often tedious and time consuming.However, with the improvements of position estimation and object detection on phones,a lightweight and low-cost solution could become the standard for the implementationphase of an IPS. Apple recently included a Light Detection And Ranging (LiDAR) sensorin their phones, greatly improving the phones depth measurements and depth understanding.This allows for a more accurate virtual representation of an environment. This thesisstudies the accuracy of ARKit’s reconstructed world and how different environments impactthe accuracy. The thesis also investigates the use of reference points as a tool to map thereconstructed environment to a geo-referenced map, such as Google Maps and Open StreetMap. The results show that ARKit can create virtual representations with centimetre levelaccuracy for small to medium sized environments. For larger or vertical environments,such as corridors or staircases, ARKit’s SLAM algorithm no longer recognizes previouslyvisited areas, causing both duplicated virtual environments and large drift errors. With theuse of multiple reference points, we showed that ARKit can and should be considered asa viable tool for scanning and mapping small scale environments to geo-referenced floorplans.