Discrimination of viable and empty seeds of Pinus patula Schiede & Deppe with near-infrared spectroscopy

Abstract

Sustainable forest production demands a continuous supply of high quality seeds for the production of seedlings in the nursery or for direct sowing. Here, we demonstrated the potential of near infrared spectroscopy as a rapid technique to discriminate viable and empty seeds of Pinus patula Schiede & Deppe. Near infrared spectra were collected from single seeds in transmittance and reflectance modes. To discriminate viable and empty seeds, multivariate classification models were developed with partial least squares (PLS) regression using the digitized spectra as a regressor and a y-vector of artificial values (1 for viable and −1 for empty seeds) as a regressand. Viable and empty seeds were perfectly distinguished by PLS models computed on full and selected transmittance spectroscopy data, while those derived from ‘full’ NIR reflectance spectra recognized 96 % of viable and 88 % of empty seeds. Analyses made on selected NIR reflectance spectra improved the classification rate of empty seeds to 100%. Difference spectra and PLS weights indicated that the origin of spectral differences between viable and empty seeds was attributed to differences in fatty acids and proteins that were totally absent in empty seeds. The result shows the prospect of developing rapid filter-based sorting equipment that can easily be automated.