Spectral Reflectance Processing via Local Wavelength-Direction Correlations

Abstract

The spectral bidirectional reflectance distribution function (BRDF) maps incident radiation of a surface to its outgoing counterpart at different wavelengths. This function plays a fundamental role in characterizing the various types of earth surfaces. Due to its high dimensionality, the measurements, analysis, and simulation of spectral BRDF are challenging. In this letter, we introduce a new method for processing spectral reflectance using the so-called data-adjacency, i.e., the correlation between adjacent wavelengths and viewing directions. The results show that the benefits of efficient representation, noise reduction, and analysis capability can all be integrated to the data. © 2019 IEEE.