A Hyperspectral Bidirectional Reflectance Model for Land Surface

Abstract

A hyperspectral bidirectional reflectance (HSBR) model for land surface has been developed in this work. The HSBR model includes a very diverse land surface bidirectional reflectance distribution function (BRDF) database with ~40,000 spectra. The BRDF database is saved as Ross-Li parameters, which can generate hyperspectral reflectance spectra at different sensor and solar observation geometries. The HSBR model also provides an improved method for generating hyperspectral surface reflectance using multiband satellite measurements. It is shown that the land surface reflective spectrum can be easily simulated using BRDF parameters or reflectance at few preselected wavelengths. The HSBR model is validated using the U.S. Geological Survey (USGS) vegetation database and the AVIRIS reflectance product. The simulated reflective spectra fit the measurements very well with standard deviations normally smaller than 0.01 in the unit of reflectivity. The HSBR model could be used to significantly improve the quality of the reflectance products of satellite and airborne sensors. It also plays important role for intercalibration among space-based instruments and other land surface related applications.

Development of the Ames Global Hyperspectral Synthetic Data Set: Surface Bidirectional Reflectance Distribution Function

“…Therefore, we use Equation to map the multispectral MODIS BRDF parameters ($${\boldsymbol{K}}_{\lambda }^{M}$$) into the spectral library generated in Q. Yang et al. (2020) and P. Yang et al. (2020) to estimate the corresponding hyperspectral BRDF parameters ($${\boldsymbol{K}}_{\lambda }^{H}$$).…”

“…A previous study by Q. Yang et al. (2020) selected a subset of ∼900 reflectance spectra from the Aster Spectral Library (Baldridge et al., 2009) and created a library of ∼40,000 BRDF spectra. These spectra can be conveniently and accurately represented by the leading ∼20 principal components (PCs), which allows Q. Yang et al.…”

“…These spectra can be conveniently and accurately represented by the leading ∼20 principal components (PCs), which allows Q. Yang et al. (2020) derive an efficient formula to calculate hyperspectral spectra from the MODIS BRDF parameters $${\boldsymbol{K}}_{\lambda }^{M}$$ at seven bands.…”

Early Detection of Wild Rocket Tracheofusariosis Using Hyperspectral Image-Based Machine Learning

“…This study is designed for the acquisition of top-view hyperspectral images, under conditions, even if controlled, that are very close to the real ones in which wild rocket vegetation grows at the soil line. However, the system needs to be implemented for practical applications, taking into account other limiting factors such as the distribution of plant leaves at different heights and/or stages and bidirectional reflectance distribution function that defines the scattering of the reflected radiance that may require correction [93]. The research conducted here aims to support farmers towards the effective monitoring and management of tracheofusariosis of wild rocket by benefiting from the non-destructive early identification of the plant's spectral response to pathogenesis.…”

Simultaneous Retrieval of Soil, Leaf, and Canopy Parameters from Sentinel-3 OLCI and SLSTR Multi-spectral Top-of-Canopy Reflectances

“…This approach is motivated by the observation that normalised BRDF kernel factors for 168 bare soils show substantial correlation across bands (e.g [21], [22]). We eliminate the 169 spectral dependency of the kernel factors by normalising them with the spectral BHR.…”