UAV Imaging of the Riverbed in Small Tree-Lined Streams: Importance of the Light Environment

UAVs are an ideal platform for remote sensing of riverbeds in small, tree-lined streams, allowing unobstructed viewing of the channel at high spatial resolution. However, effective UAV surveying of these riverbeds is hindered by a range of phenomena associated with the complex light environments of rivers, and small tree-lined streams in particular, including reflections of the overlying cloud layer from the water surface, sunglint on the water surface, and shadows from topography and riparian vegetation. Using RGB UAV imagery acquired from a range of small, tree-lined streams under a range of light conditions, we identify the prevalence of the main phenomena – reflections of clouds, sunglint, and shadows – that hinder the ability to remotely sense the riverbed. We characterize how large a constraint this is on the optimal imaging window. We then examine the degree to which sub-optimal light conditions may restrict this window, both within the year and within the day, across Europe. Our investigations suggest that different regions across Europe will have different priorities with regard to imaging, with surveys in northern rivers emphasizing avoiding low irradiant intensity in winter, and those in southern rivers emphasizing avoiding sunglint around midday. We use our findings to suggest a protocol for improved riverbed imaging that is specific to the light environment of the stream under investigation.