The development and internal assessment of a high-resolution, non-proprietary, stereo-photogrammetric setup for hydraulic experiments

International audienceRemote sensing of riverine gravel-beds has been shown to be fundamental to derive a theoretically driven definition of the hydraulic roughness and to understand the complex processes at the sediment-water interface. Commonly, 2D gravel-bed topography was recorded and analyzed, and only more recently technology allowed the measurement of high-resolution 3D Digital Elevation Models (DEMs). Equipment to do so is limited to Terrestrial Laser-Scanners (TLS) and proprietary stereophotogrammetric systems and associated commercial software. Obtaining 3D DEMs of the gravel-bed allows the use of advanced statistical functions of riverbed elevations, like Probability Distribution Functions (PDFs) and structure functions, to characterize the spatial and temporal structural development of the riverbed surface. The promise of quick high-resolution data acquisitions, obtained with digital close-range stereophotogrammetry, which can be employed at various locations, warrants detailed research into this area. In this paper we present the development of a high-resolution, non-proprietary stereophotogrammetric setup, to be used for hydraulic experiments aimed at gravel-bed roughness characterization. Based on the quantitative assessment of the calibration process and stereo rectification of the images, means to evaluate the reliability of the system are described. It is shown that the extraction of the internal orientation of the two cameras and the external orientation of the stereo setup, as well as the rectification of the images to epipolar geometry, are crucial steps to successfully match corresponding pixels and obtain high-quality DEMs of a gravel-bed. Finally, surface plots of the measured gravel-bed topography are presented, showing how improvements are reflected in the quality of the DEMs