Near‐Infrared Hyperspectral Imaging in Food Research

Abstract

It has been known for a very long time that food products can be studied by mid-infrared (MIR) (ca. 2500-15,000 nm) and near-infrared (NIR) (780-2500 nm) spectroscopy, as they contain the C-H, O-H, and N-H bonds that have high absor-banceintheNIRandMIRwavelengthregions.Thesebondsare present in the major constituents of all biological materials. Already during the 1960s and 1970s the pioneers of NIR spectroscopy were mainly interested in food applications, for example, soybeans [1], meat [2], oilseeds [3], and cereals [3][4][5][6][7]. It was found very early that water, fat, protein, and different carbohydrates could be quantified by NIR calibrationsforawidevariety ofagriculturalproducts,half-fabricates, and finished consumer products as presented by various authors [8][9][10][11][12]. Later, constituents such as inorganic salts [13], alcohol [14], fatty acids [15], antioxidants [16,17], and phenolic compounds [17,18] were also quantified, as were physical parameters such as kernel hardness [19][20][21][22][23], maturity [24], and sensory quality [24][25][26]. These analyses are usually done on bulk materials, from which a single NIR spectrum is obtained, as the goal is to integrate over an area that is as large as possible in order to avoid sampling errors. Most food products are inhomogeneous by nature, thus requiring integration or homogenization before bulk NIR measurements. NIR Hyperspectral ImagingThe earliest scientific imaging applications were simple black and white or color images in the visual spectroscopic range, but inspired by satellite imaging, multivariate image analysis [27] was soon becoming useful in the laboratory. Already the first satellite images included wavelengths in the NIR, in addition to visual and MIR wavelengths. Nowadays, hyperspectral imaging [28] is becoming more commonly available, providing complete spectra extending into the long-wave NIR (1100-2500 nm) for each pixel in an image (refer to Chapter 2). The simplest of these images have x and y spatial coordinates and a wavelength coordinate lambda, making a three-way array called a hypercube. This hyperspectral image or hypercube thus allows the description of differences and gradients in the sample under study. Because of this, the samples can be heterogeneous. The spectra in these hyperspectral images show localized spectral features that can be used for exploration and classification. With external information (reference data) also localized, quantitative and qualitative calibrations and predictions are feasible. NIR hyperspectral imaging (NHI) also means that sample preparation different from bulk NIR spectroscopy is needed. Grinding and other homogenization methods are useful for bulk NIR spectroscopic analysis, whereas sampling and sample preparation for NIR imaging have their own peculiarities.Raman, Infrared, and Near-Infrared Chemical Imaging Edited by Slobodan Š ašić and Yukihiro Ozaki

NIR Hyperspectral Imaging at Stellenbosch University: Detection of Pre-Germination in Whole Sorghum Grains

“…Within our research team (http://academic.sun.ac.za/foodsci/MM_Research-Group.htm) we have been investigating the application of NIR hyperspectral imaging for evaluating cereal grains since 2007. We investigated endosperm texture in whole maize kernels 4,5 and showed the presence of a previously unknown, third type of endosperm. 6 The potential to detect the presence of fungi in whole maize kernels was also demonstrated.…”

Wavelength Selection FOR Rapid Identification of Different Particle Size Fractions of Milk Powder Using Hyperspectral Imaging

“…For example, in the food industry, imaging is used for visual defect detection [3], and spectroscopy is used for compositional analysis of food products and the identification of adulterants [4]. Hyperspectral imaging has been previously used in the food industry and other applications such as remote sensing, airborne surveys, astronomy, agriculture, biomedical, mineralogy, and pharmaceuticals [5]. In the food industry, HSI has been used for qualitative assessment of products, for example, monitoring the freshness and quality of meat [6], identifying the types and varieties of cereals [7], detecting defects in fruits and vegetables [8], and exploring varieties of cheese and its quality [9].…”