Browsing by Subject "hyperspectral remote sensing"

As an alternative to complex 3D-modelling of structure, the canopy spectral invariants are a novel concept to describe the average behavior of photons in a vegetated canopy. The probabilities of canopy absorption and scattering can be summarized with only three parameters (?L, p and R): The green leaf single scattering albedo (?L) describes the wavelength-dependent probabilities of absorption for each time a photon interacts with a leaf. In the event of scattering, a photon’s probabilities of reinteraction (photon recollision probability p) and exiting the canopy in a given direction (directional escape factor R) can be described as independent of wavelength; as the size of the scattering elements is considerably larger than wavelengths in the shortwave radiation budget, p and R depend only upon the structural arrangement of the scattering elements. In this work, a recently published (2011) approach to infer remotely sensed (spaceborne) hyperspectral imagery (also referred to as imaging spectroscopy data) based on the canopy spectral invariants was tested in a case study on southern boreal forests at full leaf development. An atmospherically corrected image taken with the Hyperion imaging spectrometer aboard the National Aeronautics and Space Administration’s (NASA) Earth Observing-1 (EO-1) spacecraft was interpreted with a single reference transformed green leaf scattering albedo. Transforming of a traditionally defined leaf albedo means correcting the measurements for the effect of surface reflectance, resulting in probabilities of leaf scattering and absorption given a photon interacts with the leaf internal constituents. Utilizing such transformed albedo as reference results in reference (canopy) spectral invariants describing the relative difference between the reference and the scattering properties of (theoretical) mean leaves at the scale of inference (pixel). The results of the study are parallel to those of previously published and ongoing research: In essence, even while the individual parameters p and R depend on the reference, the ratio R/(1–p) (directional escape factor to total escape probability) was found practically independent of the selection of the reference, thus implicating a possibility to develop a physically-based algorithm to infer hyperspectral imagery in vegetated areas. Moreover, the reference (canopy) spectral invariants were found as highly applicable in retrieval of forest structural properties such as dominant forest type (broadleaved, coniferous, mixed) and a quantitative estimate of the broadleaf fraction of a forest area.