Comspect: A compact model for green vegetation reflectance spectra in the 400–900 nm wavelength range

Abstract

A compact empirical model for approximate description of green vegetation reflectance (GVR) spectra in the visible and near infrared wavelength range from 400 to 900 nm is proposed. The aim is to simplify the development of cyber­physical systems for forestry, agriculture, military, and environmental monitoring where distinguishing of artificial objects from the natural background is needed. Based on hyperspectrally measured spectra and simulations with PROSPECT­D and PROSAIL bio­optical leaf and canopy models, a compact model with only a few setup points at significant wavelength values is stated. After assigning the reference unit value to the chlorophyll­caused 670 nm minimum, only four easily understood tuning parameters will define the overall view of the GVR spectrum. Fermi­Dirac distribution like sigmoid step functions and Gaussian functions are used as building blocks to describe the most important spectrum features: flat or slanted ground level, green apex, red edge step, and infrared plateau. The fitting of the common nine wavelength­related parameters and of the four sample­dependent amplitude parameters was performed on the basis of seven data sets measured by a hyperspectral camera and compact spectrograph. As an application example, assessment of the quality of the military masking colour RAL 6031 is presented. The results obtained show that in the case of maximally compact formulation, a reasonable accuracy can be achieved even if only two parameters characterizing the relative heights of the green apex and the red edge step are used.