Submitted to: Transactions of the ASABE
Publication Type: Peer reviewed journal
Publication Acceptance Date: 5/1/2011
Publication Date: 5/1/2011
Publication URL: naldc.nal.usda.gov/download/50222/PDF
Citation: Tallada, J.G., Wicklow, D.T., Pearson, T.C., Armstrong, P.R. 2011. Detection of fungus-infected corn kernels using near-infrared reflectance spectroscopy and color imaging. Transactions of the ASABE. 54(3): 1151-1158. Interpretive Summary: Mold contamination of grain products can lead to significant economic losses because of health and production issues related to mycotoxins. Corn, in particular has significant problems because of its widespread use for human and animal feed. Detection and measurement of fungal-infected corn kernels would be advantageous to minimize or eliminate consumer risk, and for breeders in identifying hybrids resistant to mold growth. Near infrared reflectance spectroscopy (NIRS) and color imaging methods were studied to detect the type and extent of mold infection, on single corn kernels, for eight common mold species. NIRS was able to accurately identify 98% of uninfected kernels compared to 89% for color imaging. Certain mold species were identified better than others using NIRS. The imaging system was not able to identify mold species well. Both methods could discriminate between the uninfected and more heavily infected kernels but neither method was adequate for identifying lesser infected kernels. Overall, results indicate the methods could potentially be used for pre-screening of samples to determine if a more time consuming, but accurate, analysis is needed.
Technical Abstract: Contamination of grain products by fungus can lead to economic losses and is deleterious to human and livestock health. Detection and quantification of fungus-infected corn kernels would be adventitious for producers and breeders in evaluating quality and in selecting hybrids with resistance to infection. This study evaluated the performance of single-kernel near-infrared reflectance spectroscopy (NIRS) and color imaging to discriminate corn kernels infected by eight fungus species at different levels of infection. Discrimination was done according to the level of infection and the mold species. NIR spectra (904 to 1685 nm) and color images were used to develop linear and non-linear prediction models using linear discriminate analysis (LDA) and multi-layer perceptron (MLP) neural networks. NIRS was able to accurately detect 98% of the uninfected control kernels at best, compared to about 89% for the color imaging. Results for detecting all levels of infection using NIR were 89% and 79% for the uninfected control and infected kernels, respectively; color imaging was able to discriminate 75% of both the control and infected kernels. In general, there was better discrimination for control kernels than for infected kernels, and certain mold species had better classification accuracy than others when using NIR. The vision system was not able to classify mold species well. The use of principal component analysis on image data did not improve the classification results, while LDA performed almost as well as MLP models. LDA and mean centering NIR spectra gave better classification models. Compared to the results of NIR spectrometry, the classification accuracy of the color imaging system was less attractive, although the instrument has a lower cost and a higher throughput.