Aller au compte twitter

Research Departments

Planetary and Space Sciences



Multi-resolution digital elevation models acquired in the rift of Asal-Ghoubbet, Republic of Djibouti (© Sébastien Labarre)
Multi-resolution digital elevation models acquired in the rift of Asal-Ghoubbet, Republic of Djibouti (© Sébastien Labarre)

Most terrestrial and planetary surfaces are generally rough and of varying roughness. Surface roughness is a physical variable that interests many domains: meteorology, hydrogeology, agriculture, volcanology, planetology, but also defense. It is a multiscale variable: microscopic scale and mesoscopic scale from 10 μm to 1 cm (soil particles or regolith); macroscopic scale of 1 cm to 1 m (clumps of earth, rock or ice aggregates, micro-fractures or lava flows); topographic scale from 1 m to several kilometers (faults, hills, craters or mountains). Each scale modulates the intensity of visible or infrared electromagnetic radiation diffused or emitted to an observer, but their relative contribution to the diffusion or emission function of this radiation is very poorly known.


This lack of knowledge is partly due to the experimental difficulties encountered in measuring the relief of natural surfaces at different scales and to the complexity of the mathematical tools used to quantify the multi-scale nature of roughness. It is also associated with the scarcity of bidirectional radiometric measurements. The aim of the CAROLInA (Characterization of multi-scAle Roughness using OpticaL ImAgery) project is to better understand how the multi-scale roughness of natural surfaces influences their scattering function and whether it is possible to determine it by optical imaging. Three complementary approaches have been implemented:


  • 1) A theoretical study of multi-scale roughness using appropriate mathematical tools: wavelet decomposition and fractal analysis;
  • 2) BRDF modeling using two radiative transfer models: the Hapke analytical model developed and perfected since the early 1960s, and LuxRender, an unbiased biased open-ray model using physically correct algorithms;
  • 3) An experiment with a field campaign organized in the rift of Asal-Ghoubbet (Republic of Djibouti) in February 2016: acquisition of geophysical (surface roughness) and radiometric data (BRDF) sets on several sites selected for their wide variety of surface roughness and albedos. These measurements complete the 21 multi-angular Pléiades 1B images acquired on January 26, 2013 in a single pass.