Simulated hyperspectral data analysis using continuum removal: case study on leaf chlorophyll prediction

Abstract

Vegetation spectrum is a very complicated curve caused by its biochemical and biophysical properties. The development of hyperspectral technology makes it possible to obtain a continuous spectral curve from visible to shortwave infrared rather than several isolated broad bands. Inversion of biochemical information from it is an important and practically valuable research. In this paper, we compared two area-based vegetation indices, TVI and ABNC, in 3 aspects: (1) sensitivity to chlorophyll concentrations, (2) insensitivity to dry matters concentrations, and (3) insensitivity to mesophyll structure. Results show that ABNC, which is based on continuum removed spectra, is much better than TVI. Continuum removal is good at elimination of the influence of dry matters and mesophyll structure in visible region and near infrared reflectance peak with our model simulated leaf spectral data