Modeling and Device Development for Chlorophyll Estimation in Vegetation

Abstract

Accurate estimation of leaf chlorophyll level by remote sensing is a challenging problem. Such estimation is especially needed in an ecologically dangerous environment. Our goal is to develop new methods that allow estimating chlorophyll concentration using remote sensing data for multiple kinds of soil and vegetation. The estimation is based on a training data set obtained from the leaf samples collected at various points on the earth’s surface. A laboratory spectrophotometer was used to measure spectral reflectance curves in the visible and near-infrared ranges of the spectrum. The spectrometer was designed to comply with the strict measurement requirements essential for robust estimation. Optical indices related to leaf-level chlorophyll estimation were used as input data to test different modeling assumptions in open canopies where density of vegetation, soil, and chlorophyll content were separately targeted using a laboratory spectrometer. The goal of the research work is to estimate chlorophyll level based on spectrum characteristics of light reflected from the earth’s surface. We have applied pattern recognition techniques as well as linear and nonlinear regression models. Unlike previously suggested approaches, our methods use the shape of the spectral curve obtained from measuring reflected light. The numerical experiments confirmed robustness of the model using input data retrieved from an ecologically dangerous environment.