Fluvial bed surface microtopography exerts important control on the distribution of in-stream habitats, flow resistance and sediment transport. Therefore it is imperative to develop tools adequate to accurately measure fluvial surfaces at the grain scale to draw useful information on river ecology and hydraulics. Close-range stereo-photogrammetry (CLSP), using one or more digital cameras, and terrestrial laser-scanning (TLS) are presently the most commonly used remote sensing techniques to measure gravel-beds with a sufficient amount of details and accuracy to reach the grain scale.
Over the last two decades, both techniques have been employed in both laboratory and field testing and are theoretically able of measuring shallow river beds. Compared with TLS, a well-developed CLSP system has the advantage of being more easily employed in the field, due to its reduced cost and smaller size. The challenge with CLSP is the development of a stable workflow from image acquisition to accurate surface structure data. Recent laboratory investigations showed that modern high-resolution camera sensors now enable CLSP to rival and even outperform TLS in terms of data resolution and accuracy.
The objective of this paper is to document and assess the use of CLSP in a field environment measuring exposed gravel bars at the grain scale. For the assessment, ground truth objects were used to assess the accurate measurement of obtained surface structures. Information derived from the CLSP data was used to monitor grain-scale roughness, particle imbrication, grain size and orientation and the changes between gravel bars.