DRONE-DRIVEN SURFACE VELOCITY MEASUREMENTS IN NATURAL RIVERS

Abstract

Predicting the flow in rivers is an important concern regarding potential damages induced by floods. In order to assess the reliability and accuracy of numerical simulation tools, detailed field or laboratory data are needed. The available methods can be classified into intrusive and non-intrusive methods depending on the intrusion of the device in the flow, or pointwise and field methods, for data acquired at a single position in the flow or over a larger area. The use of imagery techniques as non-intrusive methods providing field data is increasing for different types of applications. This paper explores the use of drone-driven LSPIV (Large Scale Particle Image Velocimetry) to measure the surface velocity. From these data, the aim is to obtain discharge measurements in the flow. However, the application of these methods still presents numerous challenges. The case studied in the present paper concerns field measurements on eight different rivers. The results obtained using LSPIV and other more classical methods such as current meter (mechanical method) for velocity are compared. The integral of differences between the two methods has a range between 0 and 7% with a mean of 3.28 % and a standard deviation of 2.57 %.The results show promising agreement between both measurement techniques and highlight the perspectives for future field applications of the combined photogrammetry and LSPIV method.