Detection of adulterants and authenticity discrimination for coarse grain flours using NIR hyperspectral imaging

Abstract

Near‐infrared (NIR) hyperspectral imaging combined with multivariate analyses were used to execute adulterants detection and authenticity discrimination for coarse grain flours including millet, corn, and soybean. Spectral data in the range of 865–1,711 nm were first extracted and preprocessed from a total of 1,080 samples of six categories of pure and adulterated flours. Principal component analysis was employed to generate scores scatter plots for differentiating specific grouping of samples. Successive projection algorithm (SPA) and competitive adaptive reweighted sampling (CARS) were adopted to select the effective wavelengths. Then identification models were developed by partial least squares discriminant analysis (PLS‐DA) using effective wavelengths and full spectra (FS) as inputs, respectively. The results indicated all of models provided the satisfactory identification rate, over 94.8% for adulterated flours, and even achieved 100% for pure flours. Considering computational cost reduction and model simplification, PLS‐DA‐CARS and PLS‐DA‐SPA models were more effective in practice within an acceptable range in comparison with PLS‐DA‐FS model. Especially, PLS‐DA‐SPA model could achieve higher identification accuracy than PLS‐DA‐CARS model. It can be concluded that NIR hyperspectral imaging technology as a noncontact and quick analytical method has promising application prospects for quality control of flour products. Practical Applications Coarse grain flours and its products are more popular with people in recent years because they contain abundant vitamins, dietary fiber, microelements, and other nutrients. Driven by huge profits, some coarse grain flours such as millet, corn, and soybean are often sold in term of adulterated flours, which not only disrupts market order, but also results in quality degradation and great economic losses. As a noncontact and quick analytical technology, NIR hyperspectral imaging was used for detecting the adulteration of coarse grain flours to eliminate fraudulent or unfair practices. Multivariate analysis methods were employed to speed the detection process and improve operation efficiency. The results indicated that NIR hyperspectral imaging technology was feasible and effective in discriminating the authenticity of coarse grain flours, laying a foundation for designing a portable inspection device with the optical filters.

Rapid analysis technologies with chemometrics for food authenticity field: A review

“…As a special spectroscopic technique, HSI is also gradually being developed in the food authenticity field, primarily employing NIR and VIS-NIR for breeds or species identification in plant-derived products in recent years. Shao et al (2019) used NIR-HSI to identify millet, corn, and soybean binary mixtures and their pure samples. The PLS-DA results based on the selected effective wavelengths and full spectra showed that the discrimination rates of all the models exceeded 94.8%.…”

Nondestructive detection for adulteration of panax notoginseng powder based on hyperspectral imaging combined with arithmetic optimization algorithm‐support vector regression

“…Bai et al (2020) used hyperspectral imaging combined with principal component analysis and clustering analysis to detect the adulteration of sorghum, and the detection accuracy reached 91%. Shao et al (2019) used hyperspectral imaging combined with multivariate analysis to detect the adulteration of millet, corn, and soybean powder, and the detection accuracy was above 94.8%. Laborde et al (2021) successfully detected the adulteration of peanut powder in chocolate powder by using hyperspectral imaging combined with stoichiometry.…”