UAVS TO ASSESS THE EVOLUTION OF EMBRYO DUNES
- 1Engineering Department, University of Ferrara, Saragat 1, 44122, Ferrara, Italy
- 2Physics and Earth Sciences Department, University of Ferrara, Saragat 1, 44122, Ferrara, Italy
- 3ERREALCUBO, Engineering and Architecture Technical Studio, Ghirardacci 1, 40137, Bologna, Italy
Keywords: Embryo Dunes, Geomorphological Evolution, Coastal Environment, Unmanned Aerial Vehicles (UAVs), Structure from Motion (SfM), Photogrammetry, Low-Cost Multispectral Sensor, Agisoft PhotoScan
Abstract. The balance of a coastal environment is particularly complex: the continuous formation of dunes, their destruction as a result of violent storms, the growth of vegetation and the consequent growth of the dunes themselves are phenomena that significantly affect this balance. This work presents an approach to the long-term monitoring of a complex dune system by means of Unmanned Aerial Vehicles (UAVs). Four different surveys were carried out between November 2015 and November 2016. Aerial photogrammetric data were acquired during flights by a DJI Phantom 2 and a DJI Phantom 3 with cameras in a nadiral arrangement. GNSS receivers in Network Real Time Kinematic (NRTK) mode were used to frame models in the European Terrestrial Reference System. Processing of the captured images consisted in reconstruction of a three-dimensional model using the principles of Structure from Motion (SfM). Particular care was necessary due to the vegetation: filtering of the dense cloud, mainly based on slope detection, was performed to minimize this issue. Final products of the SfM approach were represented by Digital Elevation Models (DEMs) of the sandy coastal environment. Each model was validated by comparison through specially surveyed points. Other analyses were also performed, such as cross sections and computing elevation variations over time. The use of digital photogrammetry by UAVs is particularly reliable: fast acquisition of the images, reconstruction of high-density point clouds, high resolution of final elevation models, as well as flexibility, low cost and accuracy comparable with other available techniques.