Detection of aflatoxin B₁ on corn kernel surfaces using visible-near infrared spectroscopy

Abstract

In this study, visible-near infrared spectroscopy over the spectral range of 400–2500 nm was utilized to detect surface contamination of corn kernels with aflatoxin B1. The artificially contaminated samples were prepared by dropping known amounts of aflatoxin B1 standard dissolved in 50:50 ( v/ v) methanol/water solution, onto corn kernel surface to achieve different contamination levels of 10, 20, 50, 100, 500, and 1000 ppb. Both endosperm and germ sides of corn kernels were used for artificial contamination, and a total of 210 contaminated and control kernels were scanned with the visible-near infrared spectroscopy in reflectance mode. Spectral preprocessing methods including standard normal variate, first derivative, second derivative, first derivative + standard normal variate, and second derivative + standard normal variate were applied on the original absorbance spectra. Using the original and preprocessed spectra, the 3-class and 7-class discriminant models were established by the chemometric methods of principal component analysis-linear discriminant analysis and partial least squares discriminant analysis separately. The results showed that in discriminating the aflatoxin B1 contamination levels, the spectral range II (1120–2470 nm) generally performed better than using the corresponding spectra type over range I (410–1070 nm). Compared to using the original spectra, the first derivative and second derivative spectra generally improved the performance of the classification models. For classification thresholds of 20 and 100 ppb in aflatoxin B1, the best 3-class models achieved the same overall accuracy of 98.6% for prediction over both ranges I and II. For the 7-class discriminant models, the best overall accuracies obtained over ranges I and II are 91.4 and 97.1% for prediction.

Research advancements in optical imaging and spectroscopic techniques for nondestructive detection of mold infection and mycotoxins in cereal grains and nuts

“…As the samples used for the development of the model were naturally contaminated, the variation in the spectra could be due to decrease in the starch and protein content as a result of fungal invasion. Tao et al (2020) studied the feasibility of VIS-NIR spectroscopy for the detection of AF in corn kernels and observed differences in absorbance spectra between the fungus-infected and control kernels (Figure 4), which can be explained by the scattering of light due to metabolic activities of fungus in the kernel. They further investigated the application of NIR spectroscopy for the detection of surface contamination of corn kernels with aflatoxin B1 using PC-LDA and PLS-DA classification models separately.…”

Classification of Aflatoxin B1 Concentration of Single Maize Kernel Based on Near-Infrared Hyperspectral Imaging and Feature Selection

“…Table 6 shows the results of the comparison between the proposed method and other researchers’ methods. As shown in Table 6 , Tao et al [ 11 ] and Chakraborty et al [ 12 ] used PLS-DA to select 18 and 12 LVs, respectively, and established a classification model, which achieved fair classification results. Kimuli et al [ 13 ] selected 12 PCs of PCA to build their classification model, which achieved an ideal result.…”

“…It is difficult to control the content of AFB1 produced by natural mold, therefore the artificial inoculation of toxin becomes a simple and effective sample preparation method due to the surface distribution characteristics of AFB1 [ 38 ]. Tao et al [ 11 ], Chakraborty et al [ 12 ], and Kimuli et al [ 13 ] also prepared the contaminated maize kernel samples with the same or similar method as this study.…”

“…PCA and a least squares support vector machine (LS-SVM) were used to classify contaminated and healthy kernels, with the overall prediction accuracy of 95.33%. Tao et al [ 11 ] established the classification models of maize kernels with different AFB1 concentrations with 410–1070 nm and 1120–2470 nm spectra, respectively, and the results showed that the effect of using NIR spectra was better than that of VNIR bands. Chakraborty et al [ 12 ] classified maize kernels with six different concentrations of AFB1, using hyperspectral images at 400–1000 nm, and the study showed that partial least squares discriminant analysis (PLS-DA) could achieve good results under 12 Latent Variables (LVs).…”

“…Zhu et al [ 10 ] achieved high classification accuracy, but only two concentrations classification were studied; the concentration classes were too few to carry out a more detailed classification. Tao et al [ 11 ] and Chakraborty et al [ 12 ] studied the differentiation of multiple concentration levels; Kimuli et al [ 13 ] discussed the sample differences of four different regions of origin, but their research samples specific to a certain concentration are too few to be widely used.…”