Estimating soil roughness indices on a ridge-and-furrow surface using stereo photogrammetry

The paper deals with the quantitative characterization of small-scale random roughness on agricultural bare soils which controls many of the hydraulic and erosion processes on the field scale. More precisely, our aim is to analyse the adequacy of a stereo photogrammetry system to obtain accurate estimation of this random roughness by means of statistical parameters and to detect soil surface roughness changes due to rainfall. rk presented in this paper is based on a set of digital elevation models (DEMs) of actual agricultural bare soils obtained by stereo photogrammetry. The considered field surfaces correspond to various tillage practices (conventional seedbed, chisel and conventional ploughing) and are watered by simulated rainfalls in order to get various patterns. ereo photogrammetry process is carefully analysed; the effects of the correlation window size are taken into consideration in order to propose optimized DEM reconstructions. cal roughness parameters such as root mean square of the heights, correlation length and tortuosity are estimated on the DEMs of the database and results concerning the effect of the DEM size on the obtained accuracy are presented for each roughness parameter. The tortuosity comes out to be a relevant roughness estimator able to quantify the roughness evolution during rain, even with important degradation of the soil. y to study the evolution of roughness with rainfall thoroughly, we introduced two positional tortuosity values computed independently over the areas of rigdes and interrows of the DEM. The obtained values clearly show that the rainfalls do not decrease homogeneously the soil small-scale roughness: the interrows areas are much more smoothed by the rain than the ridges areas do. udy presented shows that stereo photogrammetry provide DEMs that enable accurate studies of the geometrical properties of soils that can definitely be of use for hydraulic and erosion studies.