Accuracy of UAV Photogrammetry in Glacial and Periglacial Alpine Terrain: A Comparison With Airborne and Terrestrial Datasets

Groos, Alexander R.; Aeschbacher, Reto; Fischer, Mauro; Kohler, Nadine; Mayer, Christoph; Senn-Rist, Armin (2022). Accuracy of UAV Photogrammetry in Glacial and Periglacial Alpine Terrain: A Comparison With Airborne and Terrestrial Datasets. Frontiers in Remote Sensing, 3, pp. 1-16. Frontiers 10.3389/frsen.2022.871994

Preview

Text Groos_et_al._2022_Frontiers_in_Remote_Sensing.pdf - Published Version Available under License Creative Commons: Attribution (CC-BY). Download (5MB) | Preview

Unoccupied Aerial Vehicles (UAVs) equipped with optical instruments are increasingly deployed in high mountain environments to investigate and monitor glacial and periglacial processes. The comparison and fusion of UAV data with airborne and terrestrial data offers the opportunity to analyse spatio-temporal changes in the mountains and to upscale findings from local UAV surveys to larger areas. However, due to the lack of gridded high-resolution data in alpine terrain, the specific challenges and uncertainties associated with the comparison and fusion of multi-temporal data from different platforms in this environment are not well known. Here we make use of UAV, airborne, and terrestrial data from four (peri)glacial alpine study sites with different topographic settings. The aim is to assess the accuracy of UAV photogrammetric products in complex terrain, to point out differences to other products, and to discuss best practices regarding the fusion of multi-temporal data. The surface geometry and characteristic geomorphological features of the four alpine sites are well captured by the UAV data, but the positional accuracies vary greatly. They range from 15 cm (root-mean-square error) for the smallest survey area (0.2 km2) with a high ground control point (GCP) density (40 GCPs km−2) to 135 cm for the largest survey area (>2.5 km2) with a lower GCP density (<10 GCPs km−2). Besides a small number and uneven distribution of GCPs, a low contrast, and insufficient lateral image overlap (<50–70%) seem to be the main causes for the distortions and artefacts found in the UAV data. Deficiencies both in the UAV and airborne data are the reason for horizontal deviations observed between the datasets. In steep terrain, horizontal deviations of a few decimetres may result in surface elevation change errors of several metres. An accurate co-registration and evaluation of multi-temporal UAV, airborne, and terrestrial data using tie points in stable terrain is therefore of utmost importance when it comes to the investigation of surface displacements and elevation changes in the mountains. To enhance the accuracy and quality of UAV photogrammetry, the use of UAVs equipped with multi-spectral cameras and high-precision positioning systems is recommended, especially in rugged terrain and snow-covered areas.

Interest & Impact

Downloads

118 since deposited on 14 Jun 2022
42 in the past 12 months

Citations

5 Citations in Web of Science ®
6 Citations in Scopus

Search

in Google Scholar™

Services

Actions (login required)

Edit item

Actions (login required)

Edit item