ACCURACY ASSESSMENT OF COASTAL TOPOGRAPHY DERIVED FROM UAV IMAGES

Autor: Nicolas Lachaussée, Frédéric Pouget, Nathalie Long, B. Millescamps, Xavier Bertin, A. Dumon
Přispěvatelé: LIttoral ENvironnement et Sociétés - UMRi 7266 (LIENSs), Université de La Rochelle (ULR)-Centre National de la Recherche Scientifique (CNRS)
Rok vydání: 2016
Předmět:
Zdroj: The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XLI-B1, Pp 1127-1134 (2016)
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Copernicus GmbH (Copernicus Publications), 2016, ⟨10.5194/isprsarchives-XLI-B1-1127-2016⟩
Xxiii Isprs Congress, Commission I (1682-1750) (Copernicus Gesellschaft Mbh), 2016, Vol. 41, N. B1, P. 1127-1134
ISSN: 2194-9034
1682-1750
DOI: 10.5194/isprs-archives-xli-b1-1127-2016
Popis: To monitor coastal environments, Unmanned Aerial Vehicle (UAV) is a low-cost and easy to use solution to enable data acquisition with high temporal frequency and spatial resolution. Compared to Light Detection And Ranging (LiDAR) or Terrestrial Laser Scanning (TLS), this solution produces Digital Surface Model (DSM) with a similar accuracy. To evaluate the DSM accuracy on a coastal environment, a campaign was carried out with a flying wing (eBee) combined with a digital camera. Using the Photoscan software and the photogrammetry process (Structure From Motion algorithm), a DSM and an orthomosaic were produced. Compared to GNSS surveys, the DSM accuracy is estimated. Two parameters are tested: the influence of the methodology (number and distribution of Ground Control Points, GCPs) and the influence of spatial image resolution (4.6 cm vs 2 cm). The results show that this solution is able to reproduce the topography of a coastal area with a high vertical accuracy (< 10 cm). The georeferencing of the DSM require a homogeneous distribution and a large number of GCPs. The accuracy is correlated with the number of GCPs (use 19 GCPs instead of 10 allows to reduce the difference of 4 cm); the required accuracy should be dependant of the research problematic. Last, in this particular environment, the presence of very small water surfaces on the sand bank does not allow to improve the accuracy when the spatial resolution of images is decreased.
Databáze: OpenAIRE