Feasibility of detecting aflatoxin B1 on inoculated maize kernels surface using Vis/NIR hyperspectral imaging

Authors: WANG, WEI - China Agricultural University; Heitschmidt, Gerald - Jerry; Windham, William; Feldner, Peggy; Ni, Xinzhi; CHU, XUAN - China Agricultural University

Submitted to: Journal of Food Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 10/29/2014
Publication Date: 1/15/2015
Citation: Wang, W., Heitschmidt, G.W., Windham, W.R., Feldner, P.W., Ni, X., Chu, X. 2015. Feasibility of detecting aflatoxin B1 on inoculated maize kernels surface using Vis/NIR hyperspectral imaging. Journal of Food Science. 80:M116-M122.

Interpretive Summary: Feasibility of using a visible/near-infrared hyperspectral imaging system to detect and differentiate different levels of aflatoxin on artificially-contaminated maize kernel surface was examined. To reduce the color effect of maize kernels, image analysis was limited to a subset of original spectra (600 to 1000 nm). Residual staining from the aflatoxin on kernel surface was selected as the region of interest for image analysis. Principal component analysis and a stepwise factorial discriminant analysis were performed on the hyperspectral image data. The results indicated that one of the discriminant factors can be used to separate control samples from all other groups of aflatoxin-contaminated kernels, whereas the other group of discriminant factors can be used to identify maize kernels contaminated with aflatoxin levels as low as 10 ppb. The overall classification accuracy was 98%. Further analysis of coefficients of the peaks identified several key wavelengths to differentiate maize kernels with or without AFB1, as well as those with different levels of aflatoxin contamination. The experiment demonstrated that that Vis/NIR hyperspectral imaging technology combined with the appropriate statistical analyses was a practical method to detect and differentiate different levels of aflatoxin on maize kernels with artificial contaminations. However, the potential to detect and differentiate naturally-occurring aflatoxin contaminations in maize kernels from the field needs to be further examined.

Technical Abstract: The feasibility of using a visible/near-infrared hyperspectral imaging system with a wavelength range between 400 and 1000 nm to detect and differentiate different levels of aflatoxin B1 (AFB1) artificially titrated on maize kernel surface was examined. To reduce the color effects of maize kernels, image analysis was limited to a subset (600 to 1000 nm) of original spectra. Residual staining from the AFB1 on the kernel surface was selected as the region of interest for analysis. Principal component analysis (PCA) was applied to reduce the dimensionality of hyperspectral image data, and then a stepwise factorial discriminant analysis (FDA) was performed on latent PCA variables. The results indicated that discriminant factors F2 can be used to separate control samples from all other groups of kernels with AFB1 contaminations, whereas the discriminant factors F1 can be used to identify maize kernels with levels of AFB1 as low as 10 ppb. The overall classification accuracy was 98%. Finally, the peaks of ß coefficients of the discriminant factors F1 and F2 were examined and several key wavelengths were identified for differentiating maize kernels with and without AFB1, as well as those with different levels of AFB1 contamination. Results indicated that Vis/NIR hyperspectral imaging technology combined with the PCA–FDA was a practical method to detect and differentiate different levels of AFB1 with artificial contamination on maize kernel surface. However, the potential application to detect and differentiate naturally-occurring toxins in maize kernels from the field still needs to be further examined.