ABSTRACT
Developments in miniaturised electronic components and Structure-from-Motion (SfM) photogrammetry software have led to the widespread use of Unmanned Aerial Systems (UASs) for multiple applications including topographical surveying. To date, much discussion around the accuracy UAS can achieve has been focused on data acquisition techniques and photogrammetric processing techniques. These factors are critical; however, this chapter aims to investigate the effects of ground control point (GCP) frequency and spatial distribution on Root Mean Square Error (RMSE) accuracy values. Three multi-rotor UASs were used to collect three aerial image datasets at two locations. The aerial images were processed using Pix4D Mapper Pro, where checkpoints were used to determine the effects of changes in the frequency and spatial location of GCPs. Additionally, qProf – a plugin for Quantum GIS (QGIS) – was used to analyse elevational differences in a variety of generated Digital Surface Models (DSMs). Information deducted from the tests indicated that although there was a strong correlation between a reduction in GCPs and an increase in RMSE, other factors including the proximity to GCPs, their spatial distribution, Ground Sampling Distance (GSD), and site topography all have a significant influence on RMSE values.
