Effects of Different Illumination and Observation Techniques of Cultivated Soils on Their Hyperspectral Bidirectional Measurements Under Field and Laboratory Conditions

This paper evaluates the fitting of the hyperspectral bidirectional reflectance data of soil surfaces formed by a cultivator, a pulverizing harrow, and a smoothing harrow, collected in field conditions as illuminated by direct and diffuse solar radiation, to their bidirectional reflectance equivalents measured in the laboratory with only a direct radiation component using the same soil materials shaped such that their roughness was similar to that formed in the field by the farming tools mentioned above. Both kinds of data were measured by the same ASD FieldSpec 3 spectrophotometer attached to goniometric devices, which observed the soil surfaces under the same directions, pointing at various fragments of the surface under field conditions and always pointing at the center of the soil samples under laboratory conditions. The worst average fit for the analyzed soil surfaces did not exceed 36%. The fit was weaker if the soil was spectrally darker with a lower spectrum level, especially at lower solar zenith angles and higher soil surface roughness. It was found that the fit increased from 400 to 450 nm, and decreased especially for wavelengths between 1950 and 2300 nm. A less significant decrease in fit was revealed at around of 700, 940, and 1140 nm.