Vincent Baeten

Oil and Fat Classification by Selected Bands of Near-Infrared Spectroscopy

One hundred and four edible oil and fat samples from 18 different sources, either vegetable (Brazil nut, coconut, corn, sunflower, walnut, virgin olive, peanut, palm, canola, soybean, sunflower) or animal (tallow and hydrogenated fish), have been analyzed by high-performance gas chromatography (HPGC) and near-infrared spectroscopy (NIRS). Fatty acids were quantified by HPGC. The near-infrared spectral features of the most noteworthy bands were studied and discussed to design a filter-type NIR instrument. An arborescent structure, based on stepwise linear discriminant analysis (SLDA), was built to classify the samples according to their sources. Seven discriminant functions permitted a successive discrimination of saturated fats, corn, soybean, sunflower, canola, peanut, high oleic sunflower, and virgin olive oils. The discriminant functions were based on the absorbance values, between three and five, from the 1700–1800 and 2100–2400 nm regions. Chemical explanations are given in support of the selected wavelengths. The arborescent structure was then checked with a test set, and 90% of the samples were correctly classified.

Multivariate calibration and chemometrics for near infrared spectroscopy: which method?

The four most important regression methods are evaluated on very large data sets: Multiple Linear Regression (MLR), Partial Least Squares (PLS), Artificial Neural Network (ANN) and a new concept called “LOCAL” (PLS with selection of a calibration sample subset of the closest neighbours for each sample to predict). The Standard Errors of Prediction (SEPs) are statistically tested and the results show that the regression methods are almost equal and that the data matrices are more important than the fitting methods themselves. The types of pre-treatments (Multiplicative Scatter Correction, Detrend, Standard Normal Variate, derivative etc.) of the spectra are too numerous to be able to test all the combinations. For each test, the pre-treatment found as the best with the PLS method is fixed for the other ones. The second part of the paper emphasises the importance of the number of samples. If any agricultural commodity, and probably any kind of product measured by an NIR instrument, can be considered as a mixture of several constituents, the databases built by collecting actual samples bringing new information can reach hundreds, if not thousands, of samples.

Hyperspectral Imaging Applications in Agriculture and Agro-Food Product Quality and Safety Control: A Review

Abstract In this review, various applications of near-infrared hyperspectral imaging (NIR-HSI) in agriculture and in the quality control of agro-food products are presented. NIR-HSI is an emerging technique that combines classical NIR spectroscopy and imaging techniques in order to simultaneously obtain spectral and spatial information from a field or a sample. The technique is nondestructive, nonpolluting, fast, and relatively inexpensive per analysis. Currently, its applications in agriculture include vegetation mapping, crop disease, stress and yield detection, component identification in plants, and detection of impurities. There is growing interest in HSI for safety and quality assessments of agro-food products. The applications have been classified from the level of satellite images to the macroscopic or molecular level.

A Backward Variable Selection method for PLS regression (BVSPLS).

Variable selection has been discussed in many papers and it became an important topic in areas as chemometrics and science in general. Here a backward iterative step-by-step wrapper method is proposed using PLS. The root-mean-square error of prediction (RMSEP) for an independent test set is used as selection criterion to quantify the gain obtained using the selected range of variables. The method has been applied to different data sets and the results obtained revealed that one can improve or at least keep constant the prediction performances of the PLS models compared to the full-spectrum models. Moreover with the advantage that the number of variables is reduced driving to an easier interpretation of the relationship between model and sample composition and/or properties. The aim is not to compare to other variable selection methods but to show that a simple one can improve or at least keep constant the prediction performances of the PLS models by using only a limited number of variables.

Non-destructive measurement of vitamin C, total polyphenol and sugar content in apples using near-infrared spectroscopy.

BACKGROUND The vitamin C and polyphenol content of apples constitute quality and nutritional parameters of great interest for consumers, especially in terms of health. They are commonly measured using laborious reference methods. The purpose of this study was to evaluate the potential of near-infrared (NIR) spectroscopy as a rapid and non-destructive method to determine the sugar, vitamin C and total polyphenol content in apples. RESULTS The quality parameters of more than 150 apple genotypes were analyzed using NIR and reference methods. The results obtained using the least squares support vector machine regression method showed good to very good prediction performance. Low standard error of prediction values, in addition to relatively high ratio to prediction (RPD) values, demonstrated the precision of the models, especially for polyphenol and sugar content. High RPD values (5.1 and 4.3 for polyphenol and sugar, respectively) indicated that an accurate classification of apples based on their content could be achieved. CONCLUSION NIR spectroscopy coupled with the multivariate calibration technique could be used to accurately measure the quality parameters of apples. In addition, in the case of breeding programs, NIR spectroscopy can be considered an interesting tool for sorting varieties according to a range of concentrations.

Discrimination of fish bones from other animal bones in the sedimented fraction of compound feeds by near infrared microscopy

Since the bovine spongiform encephalopathy (BSE) crisis, the use of animal proteins in animal feed has been prohibited. From October 2003, the European Union (EU) adopted Regulation (EC) no. 1774/2002 governing animal by-products (ABPs), which seeks to address the possible risks inherent in recycling potential infectivity due to the absence of barriers within species and to exclude the cannibalism which may be induced by intra-species recycling. There is an urgent need to develop fast and reliable methods for identification of low-level ABP origins. In this study, near infrared (NIR) microscopy was used to identify different classes of ABPs. Samples of fish meals (n = 10) and meals of land-animal origin (n = 50) were ground, sedimented and analysed using an Auto Image Microscope connected to a Fourier transform near infrared spectrometer (FT-NIR). Sediment fraction particles were spread on a Spectralon plate, presented to the NIR microscope and scanned in the 1112–2500 nm region. The support vector machine (SVM) algorithm was used to construct models to identify class origin. Models correctly classified 100% of the samples in the calibration set and between 95 and 95.5% in the validation set. The results demonstrated the potential of FT-NIR microscopy as a rapid method for distinguishing between fish and land-animal particles.

Near infrared spectroscopy (NIRS) for on-line determination of quality parameters in intact olives

The acidity, moisture and fat content in intact olive fruits were determined on-line using a NIR diode array instrument, operating on a conveyor belt. Four sets of calibrations models were obtained by means of different combinations from samples collected during 2009-2010 and 2010-2011, using full-cross and external validation. Several preprocessing treatments such as derivatives and scatter correction were investigated by using the root mean square error of cross-validation (RMSECV) and prediction (RMSEP), as control parameters. The results obtained showed RMSECV values of 2.54-3.26 for moisture, 2.35-2.71 for fat content and 2.50-3.26 for acidity parameters, depending on the calibration model developed. Calibrations for moisture, fat content and acidity gave residual predictive deviation (RPD) values of 2.76, 2.37 and 1.60, respectively. Although, it is concluded that the on-line NIRS prediction results were acceptable for the three parameters measured in intact olive samples in movement, the models developed must be improved in order to increase their accuracy before final NIRS implementation at mills.

Evaluation of the overall quality of olive oil using fluorescence spectroscopy.

The fluorescence spectra of some olive oils were examined in their natural and oxidised state, with wavelength range emissions of 300-800 nm and 300-400 nm used as excitation radiation. The fluorescence emissions were measured and an assessment was made of the relationship between them and the main quality parameters of olive oils, such as peroxide value, K232, K270 and acidity. These quality parameters (peroxide value, K232, K270 and acidity) are determined by laboratory methods, which though not too sophisticated, they are required solvents and materials as well as time consuming and sample preparation; there is a need for rapid analytical techniques and a low-cost technology for olive oil quality control. The oxidised oils studied had a strong fluorescence band at 430-450 nm. Extra virgin olive oil gave a different but interesting fluorescence spectrum, composed of three bands: one low intensity doublet at 440 and 455 nm; one strong band at 525 nm; and one of medium intensity at 681 nm. The band at 681 nm was identified as the chlorophyll band. The band at 525 nm was derived, at least partially, from vitamin E. The results presented demonstrate the ability of the fluorescence technique, combined with multivariate analysis, to characterise olive oils on the basis of all the quality parameters studied. Prediction models were obtained using various methods, such as partial least squares (PLS), N-way PLS (N-PLS) and external validation, in order to obtain an overall evaluation of oil quality. The best results were obtained for predicting K270 with a root mean square (RMS) prediction error of 0.08 and a correlation coefficient obtained with the external validation of 0.924. Fluorescence spectroscopy facilitates the detection of virgin olive oils obtained from defective or poorly maintained fruits (high acidity), fruits that are highly degraded in the early stages (with a high peroxide value) and oils in advanced stages of oxidation, with secondary oxidation compounds (high K232 and K270). The results indicate the potential of a spectrofluorimetric method combined with multivariate analysis to differentiate, and even quantify, the levels of oil quality. The proposed methodology could be used to accelerate analysis, is inexpensive and allows a comprehensive assessment to be made of olive oil quality.

Calibration transfer from dispersive instruments to handheld spectrometers.

The Foss NIRSystem 6500 is one of the most commonly used laboratory instruments in agriculture and in particular in feed. New technological developments include micro-electro mechanical system (MEMS) technology, used in miniature handheld instruments such as the Polychromix Phazir spectrometer that are increasingly required for on-site analysis. The objective of this study was to assess the potential of a calibration transfer from the Foss NIRSystem 6500 to the Polychromix Phazir. The results show that good calibration models were obtained for various feed properties (fat, fiber, protein, and starch) developed on a Foss NIRSystem 6500, based on a spectral database of 9164 samples transferred to a Polychromix Phazir handheld spectrometer.

An overview of the legislation and light microscopy for detection of processed animal proteins in feeds

From the first cases of bovine spongiform encephalopathy (BSE) among cattle in the United Kingdom in 1986, the route of infection of BSE is generally believed by means of feeds containing low level of processed animal proteins (PAPs). Therefore, many feed bans and alternative and complementary techniques were resulted for the BSE safeguards in the world. Now the feed bans are expected to develop into a “species to species” ban, which requires the corresponding species‐specific identification methods. Currently, banned PAPs can be detected by various methods as light microscopy, polymerase chain reaction, enzyme‐linked immunosorbent assay, near infrared spectroscopy, and near infrared microscopy. Light microscopy as described in the recent Commission Regulation EC/152/2009 is the only official method for the detection and characterization of PAPs in feed in the European Union. It is able to detect the presence of constituents of animal origin in feed at the level of 1 g/kg with hardly any false negative. Nevertheless, light microscopy has the limitation of lack of species specificity. This article presents a review of legislations on the use of PAPs in feedstuff, the detection details of animal proteins by light microscopy, and also presents and discusses the analysis procedure and expected development of the technique. Microsc. Res. Tech., 2011.