Performing low-altitude photogrammetric surveys, a comparative analysis of user-grade unmanned aircraft systems

Abstract

This paper shows results of comparing performances of four unmanned aircraft systems (UAS) in terms of photogrammetric survey’s quality. This study aims to investigate what is the more suitable UAS for specific applications considering the required scale factor, such as for architectural, environmental, and restoration purposes. A series of photogrammetric surveys were conducted in a hilly area of about 5 ha using Phantom 4 Adv, Mavic 2 Pro, Mavic Air 2, and Mavic Mini 2. These unmanned aircrafts are commercial user–grade systems used mainly by private professionals. Several photogrammetric reconstructions were performed by varying essential parameters, such as flight altitude and cameras of remotely piloted aircraft systems (RPAS), applying structure-from-motion (SfM) algorithms to the images taken from the UAS. The surveys’ quality was analyzed by comparing the ground targets’ coordinates extrapolated from the point clouds to those measured on the field with indirect georeferencing through GNSS technology. Fifty targets were installed and arranged following a reasonably regular mesh. The boundary conditions were maintained the same for each flight mission, flight trajectories, and the ground control point distribution on the ground. For each survey made by each of the four UAS, altimetric and planimetric residuals were reported and compared. Average residuals from Phantom 4 Adv, about 15 mm, almost disappear compared to the other UASs; the discrepancy is one order of magnitude. With a regular grid geometry of ground targets, the Mavic Mini 2 led to an error average of about 5 cm. Remembering that the Mavic Mini 2 is an ultralight drone (does not require a pilot's license), it could significantly reduce cost compared to the other systems.