JNIRS RSS Feed

Editorial

No Abstract Available

How often do references need to be measured when using a near infrared diode array spectrometer

The stability of a diode array spectrometer in the near infrared region has been investigated in two experiments where the frequency of passing of dark and white references, which need manual intervention, was varied. In the first experiment, an initial pair of references was used to standardise spectra taken over the next four days. In the second, references were taken at hourly intervals over a period of ten hours. The conclusion is that, in a reasonably well-controlled environment, the spectrometer is stable over long periods and the passing of hourly references conveys no advantage. An important implication is that this spectrometer may be used in on-line applications without the need to construct an automatic mechanism to measure references.

Prediction of the chemical composition of poultry excreta by near infrared spectroscopy

The potential of near infrared (NIR) spectroscopy for the determination of the chemical composition of poultry excreta was investigated, within the framework of studies on heritability of digestive efficiency in broilers. Samples in the calibration and validation databases (DB1 and DB2) corresponded to animals fed with a similar wheat-based diet. A second validation study was performed on excreta samples from animals fed more variable diets, including peas and maize (DB3). Excreta samples were freeze-dried and ground. Near infrared reflectance spectra were taken on a monochromator spectrometer between 400 nm and 2500 nm. Samples were analysed for mineral matter (MM), gross energy (GE), starch, crude fat (CFAT), total nitrogen (NTOT), uric acid nitrogen (NUA) and protein nitrogen estimated directly (PNTERP) or by difference between NTOT and NUA (PNUA). Depending on the parameters studied, 250 to 700 samples were analysed by reference methods. The standard error of cross-validation (SECV) and R² of calibrations were: 0.60% and 0.96 for MM, 166 kJ kg^–1 and 0.99 for GE, 0.59% and 1.00 for starch, 0.44% and 0.99 for CFAT, 0.25 and 0.89 for NTOT and 0.22 and 0.97 for NUA, respectively. Calibration for PNTERP (SECV = 0.07; R²=0.98) was much more precise than PNUA (SECV = 0.21%, R² = 0.85). Validation carried out on databases DB2 and DB3 resulted in standard errors of prediction (on DB2) and extrapolation (on DB3) generally higher than SECV, while remaining relatively precise with prediction r² values from 0.83 to 0.99 and extrapolation r² from 0.86 to 0.98, with the exception of PNUA for which r² was 0.22 and 0.64, respectively. For some parameters, the lower validation performance was due to biases, particularly in the case CFAT and NUA for prediction and MM, GE and NUA for extrapolation. Global calibrations made with DB1+DB2+DB3 were more precise (GE, NTOT) or equally precise (all other parameters) than with DB1 alone. These results confirmed the potential precision of calibrations for the major organic compounds in poultry excreta and suggested that their use could be extended to excreta issued from a wider range of diets without losing precision.

Quantitative determination of quality parameters and authentication of vodka using near ...

The objective of this study was to determine the potential of using near infrared (NIR) transmission spectroscopy to build calibration models for the quantitative characterisation and qualitative discrimination of Russian and non-Russian (foreign) vodkas. The results of partial least squares models based on the NIR spectra of 109 vodka samples showed that the major constituent alcoholic strength [root mean square error of prediction (RMSEP) 0.25% vol] and the physical parameter relative density (RMSEP 0.0003) could be successfully determined quantitatively. The method failed, however, in quantifying certain minor components such as anions, cations and sugars. For qualitative discrimination, soft independent modelling of class analogy analysis (SIMCA) and linear discriminant analysis (LDA) were applied to the sample set containing both the Russian and the foreign vodkas. Despite the correct assignment of unknown test samples to the respective vodka species, both modelling approaches, however, did not prove reliable enough for unambiguous authentication purposes.

Indirect detection of Fusarium verticillioides in maize (Zea maize L. ...

Near infrared (NIR) hyperspectral imaging and hyperspectral image analysis were evaluated for their potential to distinguish between Fusarium verticillioides infected and sound whole maize (Zea maize L.) kernels. Hyperspectral images of infected and sound kernels were acquired using a Spectral Dimensions MatrixNIR camera with a spectral range of 960–1662 nm and a sisuChema hyperspectral pushbroom imaging system with a spectral range of 1000–2498 nm. Background, bad pixels and shading were removed using exploratory principal component analysis (PCA) on absorbance images. PCA could be used effectively on the cleaned images to identify classes including infected and non-infected regions on individual kernels. A distinct difference between infected and sound kernels along principal component (PC) one with two distinguishable clusters was found. The loading line plot of the first PC of the sisuChema hypercube showed important absorbance peaks for the two classes, i.e. 1960 nm and 2100 nm for the infected class and 1450 nm, 2300 nm and 2350 nm for the non-infected class. Partial least squares discriminant analysis (PLS-DA) was applied. The coefficient of determination was 0.73 for the MatrixNIR image and 0.86 for the sisuChema image, both after three PLS components. These PLS-DA models could be used to calculate predictions on a test set image. The predictions could be shown as prediction images and an acceptable root mean square error of prediction was obtained (0.23). NIR hyperspectral imaging has the potential to be used as a rapid, objective means of indentifying fungal infected maize kernels and infected regions.

Using the frequency components of near infrared spectra: optimising calibration and ...

The modification of frequency components (Fourier coefficients and wavelet detail component) of near infrared spectra for the optimisation of calibration and standardisation processes was investigated. High-frequency components were smoothed and approximated to remove components most likely to represent noise and background information. Savitzky–Golay smoothing and signal correction were used for that purpose. Frequency modification methods were used in addition to wavelength domain processing techniques. Whole soybean protein and oil calibrations were developed on four instruments with their own calibration set (two Foss Infratecs and two Bruins OmegAnalyzerGs). A validation strategy with two sample sets of known and of new variability was implemented. Frequency modification methods showed improvements of the prediction precision in calibration [relative predictive determinant (RPD) increased from 8.57 to 9.25 for protein and from 7.01 to 7.28 for oil with Fourier coefficients-based smoothing for Infratec 12410350]. Frequency based pre-processing methods were also successful when transferring prediction models in intra and inter-brand situations (RPD of the secondary unit of 9.21 compared to original RPD of 8.45 in intra-brand for Foss network for protein; RPD of secondary unit of 9.33 compared to original RPD of 8.74 for inter-brand scenario with Foss Infratec 1241 master of Bruins units for protein). The smoothing of Fourier coefficients showed the best results. Prediction accuracies were not modified by the frequency-based modifications, except in the inter-brand scenario. An appropriate pre-processing limited the need for other standardisation methods except in inter-brand situations where a bias correction should be implemented. Frequency-based pre-treatment methods tend to specialise the calibration set to optimise predictions. This may not be suitable when the variability of the future samples is not included in the calibration set (i.e. yearly variability of agricultural products).

Inverse, classical, empirical and nonparametric calibrations in a Bayesian framework

The calibration paradigm for near infrared (NIR) spectroscopy is explored in a Bayesian framework in which a model for the dependence of the NIR spectrum on the composition of the sample is combined with a prior distribution representing beliefs about the composition of the sample to be predicted. With appropriate choices for this prior distribution it is possible to reproduce standard regression results, remove the shrinkage to the mean that is sometimes seen as a problem in inverse regression, or improve predictions by taking account of non-standard distributions of composition in the population of interest. These options are illustrated by applying them to the prediction of wheat and sunflower content using a database of 7532 commercial animal feed samples which has been studied extensively in the past. Various options for the modelling of the relationship between spectra and composition are investigated, including linear and nonlinear regressions and a nonparametric approach based on kernels. This Bayesian kernel approach, which has features in common with local calibration methods, gave results better than anything previously achieved with the animal feed database.

Quantification of absorber through a scattering medium of different thickness using evanescent ...

A non-invasive method has been developed for analyte quantification in fluids surrounded by optically scattering, opaque walls. This method is based on steady state, visible wavelength reflectance measurements made simultaneously at multiple positions on the surface of a sample. Previous work has shown that reflectance measurements contain information about underlying scattering layers in layered scattering samples. We hypothesize that similar information about an absorbing layer below a scattering layer can be obtained from evanescent wave effects. Principal component analysis showed the data to be composed of three components, which were refined by a multivariate curve resolution alternating least squares (MCR-ALS) approach with non-negativity constraints. The first component is related to the scattering layer thickness, the second is associated with analyte concentration, and the third is due to a minor back reflection within the sample cell. Both MCR and stagewise multilinear regression (SMLR) approaches were taken to estimate analyte concentration and scattering layer thickness, for samples having thicknesses between 1 and 8 mm. Results demonstrate that a simple experimental configuration can easily predict optical properties of unknown samples. With the adoption of a multi- wavelength approach to this method, it is expected that improved absorption coefficient (µa) estimation accuracy can be realized in a variety of application areas such as in analysis through opaque containers, in-vivo measurements, and in-line monitoring of reactions.

Identifying counterfeit medicines using near infrared spectroscopy

Counterfeit medicines are a growing threat to public health across the world and screening methods are needed to allow their rapid identification. A counterfeiter must duplicate both the physical characteristics and the chemical content of a proprietary product to avoid it being detected as a counterfeit product and this is almost impossible to get right. Counterfeit proprietary medicines are, therefore, relatively easy to identify by near infrared (NIR) spectroscopy which can detect physical as well as chemical differences between products by simple spectral comparison. Identifying generic products is more difficult as they use different excipients in the tablet or capsule matrix. Nevertheless, using appropriate models and a large library, NIR spectroscopy can detect counterfeit generic versions. Detecting sub-standard proprietary medicines can be carried out with NIR spectroscopy models and the most widely used is partial least squares regression (PLSR). General rules for generating accurate quantitative models are easy to describe. Quantifying the active pharmaceutical ingredient (API) in generic products can also be carried out using PLSR models with calibration samples generated by manufacturing laboratory samples or by collecting many generic versions of a medicine so as to obtain a good range of the API content in tablets and capsules. Using hand-held instruments or mobile laboratories allows NIR spectrometers to be taken to places where analyses may be made quickly, rather than taking the samples to a laboratory. This has the enormous advantage that the screening of large numbers of samples may be made in pharmacies and wholesalers. Imaging can bring a whole new dimension to NIR spectroscopy to allow the identification of the API and individual excipients as well as measuring the particle sizes of components and giving a measure of the homogeneity of the matrix. The effect of water on potential misidentifications may be obviated by only using blister-packed samples, having large spectral libraries subjected to different humidities or omitting the spectral region where water absorbs.