Litcius/Paper detail

Hyperspectral indices developed from the low order fractional derivative spectra can capture leaf dry matter content across a variety of species better

Jia Jin, Quan Wang

2022Agricultural and Forest Meteorology24 citationsDOIOpen Access PDF

Abstract

Leaf mass per area (LMA) is an important indicator of plant functioning and photosynthetic capacity and is critical for understanding plant physiology and ecosystem function. Despite detailed and continuous spectral information offered in hyperspectral reflectance, LMA remains a difficult leaf characteristic to be retrieved due to its complex constituents and overlapping absorptions with leaf water. Traditional derivative analysis is commonly used to extract the absorption band positions and to resolve overlapping spectral features, but most cases are only limited to an integral derivative that ignores the asymptotic information between spectral curves. Recent advances in fractional-order derivative (FOD) based analyses, however, have shown their advantages in eliminating background noise as well as in extracting effective information from spectral information. We have thus investigated the potentials of using the fractional derivative indices to retrieve LMA based on a composite dataset consisting of 842 leaf samples from various species. The results demonstrated that the 0.3-order FOD indices provided the highest accuracies to trace LMA and, meanwhile, had the least sensitivity to random noise. Among the ten different index types examined in this study, the SR(1320, 1715) calculated from the 0.3-order derivative spectra had the best performance with an R2 of 0.79. Furthermore, the band around 1715 nm was confirmed to be the wavelength with the highest relative absorption of LMA, while the band around 1320 nm was the non-absorbing wavelength for LMA, which could be applied as a base to describe the effects of other leaf constituents. The results of this study revealed the potential of low-order FOD indices to capture LMA and we foresee their wide applications in the future.

Topics & Concepts

Hyperspectral imagingDerivative (finance)Absorption (acoustics)Noise (video)WavelengthSecond derivativeSpectral lineSensitivity (control systems)Biological systemEnvironmental scienceMathematicsRemote sensingPhysicsOpticsComputer scienceArtificial intelligenceMathematical analysisBiologyImage (mathematics)GeologyElectronic engineeringEconomicsEngineeringAstronomyFinancial economicsRemote Sensing in AgricultureLeaf Properties and Growth MeasurementPlant Water Relations and Carbon Dynamics
Hyperspectral indices developed from the low order fractional derivative spectra can capture leaf dry matter content across a variety of species better | Litcius