Photogrammetric 3d-mapping using low-cost unmanned aerial vehicle - Analysis of accuracy

Abstract: The UAV (unmanned aerial vehicle) is a part of the digital transformation in the construction industry. With increased options of low-cost UAVs, more companies will want to integrate UAVs into their business. By capturing a large amount of data in a short time, it could be very useful for several applications within the construction industry. The question is: can the low-cost UAV be an option considering its accuracy and efficiency? This thesis will investigate how accurate a low-cost UAV can be for 3D modelling. We used a low-cost UAV and a total station in a field study in Malmö, Sweden, to obtain 3D-data from a building for further evaluation and comparison. Moreover, we analyzed the future and evaluation of the implementation of the UAV through a literature study and interviews with professionals working in the construction industry. The derived results show that it is possible to get a spatially accurate point cloud even without the use of ground control points (GCPs). This point cloud did not give accurate absolute positions to a given reference system without the use of GCPs but as that is not necessary for facade plans and other building plans, it will still be of good use. Our results show an accuracy of about 1-2 cm in a point cloud for an average residential building. Smaller features in darker areas might not be correctly modelled. It also shows that it is fairly easy to deploy a low-cost drone project for use in the construction industry due to a high level of autonomy during flight.