ABSTRACT
Hyperspectral remote sensing has been widely applied in vegetation remote sensing and has been become an important measurement tool in the study of surface vegetation thanks to its large number of narrow and contiguously spaced spectral bands. This chapter investigates the potential of hyperspectral remote sensing in estimating biological variables of cotton at the canopy level.
This study's objective was to analyze hyperspectral remote sensing capability in detecting characteristic differences in cotton crops under different nitrogen (N) application rates and different growing stages. We designed two experiments, Experiment 1 and Experiment 2, that included three N application rates, 0, 60, and 120 kg N ha-1 (termed LN, MN, and HN) and 90, 180, and 360 kg N ha-1 (termed LN, MN, and HN), respectively, and the cotton growth cycle was divided into three stages: (1) rapid growth period, (2) full green canopy, and (3) senescence period. The results were as follows:
Nitrogen increased the cotton dry weight of aboveground biomass (DWAB).
Canopy reflectance spectra under LN, MN, and HN treatments decreased successively in the NIR region, and one-way ANOVA of cotton canopy reflectance among the three N treatments at difference wavelengths showed striking differences in the NIR region.
Continuum-removed technology improved the absorption features between 550 and 750 nm, and it was demonstrated that the width and depth of absorption characteristic increased with increases in the N application rate and that the best bands for identifying nitrogen conditions were 640–690 nm.
The reflectance that was most sensitive to the N application rate was located at both sides of the chlorophyll maximal absorptance (680 nm).
The crop variables such as leaf area index, chlorophyll content, aboveground dry biomass, and seed cotton yield increased with increases in N application rates.
Overall, the results of this study suggest a way to estimate cotton canopy quality at the field level.
