Factor of bi-conical reflectance in contrasting vegetable species: modeling of zenithal angles

Abstract

The multi-angular reflectance information has been proposed for discrimination of classes of natural vegetation under different characterization approaches. Although apparently this method enables reliable definition of typical multi-angular spectral signatures, its widespread use has not been analyzed. This paper presents a model of the effects of viewing and illumination zenith angle and lighting on reflectance and is validated in a set of experiments with eight contrasting plant species plus bare ground, under bi-conical vision lighting conditions. The results show that the proposed model is robust and reliable and can thus be used operationally. Analyses to determine the effect of soil and plant density show that there are important areas of confusion (overlapping information) between species when considering combined effects for different azimuthal angles. These results challenge the widespread use of multi-angular spectral information and justify this approach only if the vegetation background and plant density (amount of foliage) are fixed.