Three-dimensional surface reconstruction by stereovision and soil roughness modeling

Three-dimensional description of soil structure has many applications in remote sensing. In active microwave remote sensing, the analysis of the backscattered radar signal by natural surfaces has shown its potential to provide information on physical parameters of the soil, especially soil moisture and roughness. To study the relationship between this signal and the physical parameters of the ground, electromagnetic models have been developed. These electromagnetic models need a good estimation of statistical parameters of the surface geometry. In this paper, a new approach to describe soil roughness is proposed. It is based on three-dimensional soil surface images obtained with fine resolution by a stereoscopic platform. The precision on estimation of statistical parameters of the reconstructed soil surface, rms height and correlation length, will be evaluated function on the geometrical cameras parameters