SURFACE FLOW VELOCITY MEASUREMENTS FROM UAV-BASED VIDEOS

Abstract. Most measuring methods for determining the volumetric flow rate or surface flow velocity have in common that they cannot be safely used under extreme outflow conditions. Especially in catastrophic situations, it is of particular interest to determine the amount of water that flows into the hinterland as precisely as possible in order to improve hydrodynamic models. Faulty assumptions lead to misleading calculations and may result in preventable casualties. As technical improvements throughout the last decade facilitated the widespread utilization of unmanned aerial vehicles (UAV) or remotely piloted aircraft systems (RPAS), these systems are now capable to collect and transmit precise information from remote areas to task forces immediately. The usage of a UAV is possible with minimal preparation at almost every place and is suitable for improving the database for a quick assessment of the status during a catastrophic event.

In this work, the determination of surface flow velocity using unmanned aerial vehicles (UAV) and floating optical tracers is evaluated. It is also discussed, to what extent numerical methods are able to efficiently undistort and correct this data. Precision analysis of video data from field investigations was performed with R using three different approaches that calculated the true velocity of the floating objects. The results indicate similar degrees of precision for both advanced methods but calculating an ortho-corrected video is a timeconsuming process not suitable for nearly real-time applications.