Non-Destructive Evaluation of Sugar content Using a Combination of Near-Infrared Spectroscopy (NIRS) and Chemometrics Methods

In this research, the ability of the reflectance near-infrared (NIR) spectrometry was investigated for non-destructive assessment of the sugar content of sugar beet roots. To this end, spectrometry of 120 samples of sugar beet was performed in the interactance measurement mode within the spectral range of 350-2500 nm using a contact probe. Spectral data obtained from the spectrophotometer included unwanted information and noise in addition to the information about the samples. In order to arrive at accurate analytical models, pre-processing of the spectral data was required prior to regression model simulation. For this purpose, multivariate calibration models of partial least squares (PLS) were developed based on the reference measurements and the information of the preprocessed spectra. A combination of different methods for assessment and prediction of sugar content was employed: smoothing, normalizing as well as increasing the spectral resolution. Prediction of the sugar content of intact samples with the PLS model based on SG D2, had the best discrimination ability. Thus, SG preprocessing (R_C^2=0.973, RMSEC = 0.306, R_P^2= 0.977, RMSEP = 0.265) is suitable for predicting beet root sugar content with high accuracy (SDR= 6.660).