A new method of unmanned aerial vehicle (UAV)-based tracer tests using RGB (red, green, blue) images was developed in order to acquire the spatio-temporal concentration distribution of tracer clouds in open channel flows. Tracer tests using Rhodamine WT were conducted to collect the RGB images using a commercial digital camera mounted on a UAV, and the concentration of Rhodamine WT using in-situ fluorometric probes. The correlation analysis showed that the in-situmeasured concentrations of Rhodamine WT were strongly correlated with the digital number (DN) of the RGB images, even though the response of DN to the concentration was spatially heterogeneous. The empirical relationship between the DN values and the Rhodamine WT concentration data was estimated using artificial neural network (ANN) models. The trained ANN models, which consider the effect of water depth and river bed, accurately retrieved the detailed spatio-temporal concentration distributions of all study areas that had an R2 higher than 0.9. The acquired spatio-temporal concentration distributions by the proposed method based on the UAV images gave general as well as detailed views of the tracer cloud moving dynamically in open channel flows that cannot be easily observed using conventional in-situ measurements.
