In an effort to address the problem of rapid detection of aflatoxin in grain, particularly oilseeds, the current study assessed the spectral differences of aflatoxin production in kernels from a cornfield inoculated with spores from 2 different strains of toxigenic Aspergillus flavus. Aflatoxin production in corn from the same field due to natural infestation was also assessed. A small corn plot in Baton Rouge, La., U.S.A., was used during the 2008-growing season. Two groups of 400 plants were inoculated with 2 different inocula and 1 group of 400 plants was designated as controls. Any contamination detected in the controls was attributed to natural infestation. A subset of each group was imaged with a visible near infra red (VNIR) hyperspectral system under ultra violet (UV) excitation and subsequently analyzed for aflatoxin using affinity column fluorometry. Group differences were statistically analyzed. Results indicate that when all the spectral data across all groups were averaged, any potential differences between groups (treated and untreated) were obscured. However, spectral analysis based on contaminated “hot” pixel classification showed a distinct spectral shift/separation between contaminated and clean ears with fluorescence peaks at 501 and 478 nm, respectively. All inoculated and naturally infected control ears had fluorescence peaks at 501 nm that differed from uninfected corn ears. Results from this study may be useful in evaluating rapid, noninvasive instrumentation and/or methodology for aflatoxin detection in grain.
The present study evaluated the possible application of fluorescence hyperspectral imaging methodology for the rapid and noninvasive detection of aflatoxin in postharvest corn during various stages of processing. Specific information from this study may be applicable when developing rapid, noninvasive instrumentation and/or methodology for whole ear inspection as well as shucked kernels.