ABSTRACT
High-resolution Digital Elevation Models (DEMs) of gravel riverbeds are common tools for studying surface roughness. However, a necessary step in these studies, the DEM quality assessment, remains challenging. To highlight its importance, we review present surface roughness characterization methods for gravel beds and propose to build a realistic gravel-bed model with known elevations using modern 3D printing technology. The model is placed in a laboratory flume and measured throughair with digital stereo photogrammetry. To evaluate the measurement performance, the measured and “truth” models are mutually aligned and compared point-by-point. Our experiments showed high accuracy, namely, the Mean Unsigned Error (MUE) of 0.43 mm and Standard Deviation of Error (SDE) of 0.62mm in the viewed area of 296×184mm2. In terms of external reliability, small discrepancies were observed between the measured and true statistical roughness parameters. DEM quality assessment with a 3D printed model will be useful for setting up the optimum strategy for gravel-bed DEM collection, both in air and through-water, in future.
