Influences of the Ground Control Point (GCP) configuration on the UAV-derived Structure from Motion (SfM) in the coastal environment

In recent years, there has been a considerable increase in the use of Unmanned Aerial Vehicles (UAVs) for conducting high-resolution topographic surveys across various environments, with excellent outcomes. Photogrammetric surveys, in particular those employing drones, constitute the primary technique for reconstructing topography. These surveys involve drone-acquired RGB images, followed by different processing phases that use the Structure from Motion (SfM) technique. Factors influencing accurate geometry reconstruction in SfM models include flat terrain, presence of water, and unidirectional model development, with coastal environments posing challenges resulting from the convergence of these factors. Ground control points (GCPs) play a crucial role in enhancing the georeferencing and geometry of the SfM models, and in aiding error estimation across three dimensions. However, it is still difficult to determine an optimal GCP density using the various practices observed in the scientific literature. The aim of this study is to investigate the GCP placement role across three distinct coastal environments, by assessing similarities and differences, and by surveying and exploiting a vast number of GCPs at varying flight heights with two drones. This work explores the role of the GCP amount and position when georeferencing the point clouds through the application of statistical approaches, and examines their influence on the DEMs. Our findings suggest that a GCP density exceeding 5 GCP/ha is necessary to obtain constant and accurate outcomes in very narrow and long beaches. Furthermore, the study proposes a control-to-check RMSE ratio for estimating model errors with limited GCPs. SfM applications in coastal settings may exhibit spatial variability in errors, emphasizing the importance of associating DEMs with error maps for accurate topographic and morphological analyses.