Skip to main content
ARS Home » Midwest Area » Peoria, Illinois » National Center for Agricultural Utilization Research » Mycotoxin Prevention and Applied Microbiology Research » Research » Research Project #446214

Research Project: Rapid Prediction and Measurement of Aflatoxin in Whole Kernel and Ground Corn Using a Portable Imager

Location: Mycotoxin Prevention and Applied Microbiology Research

Project Number: 5010-42000-052-001-N
Project Type: Non-Funded Cooperative Agreement

Start Date: Jul 15, 2024
End Date: Jul 14, 2029

Objective:
To develop models that would enable the prediction of aflatoxin content in whole kernel or ground corn using a novel handheld imaging device in combination with artificial intelligence.

Approach:
Near infrared (NIR) spectroscopy has been used for many years to measure quality parameters of grain, but because mycotoxins are frequent contaminants of grains occurring at much lower levels (i.e. part per million or part per billion levels). The use of NIR to predict mycotoxin concentrations have been only modestly successful. The cooperator has developed a technique that combines hyperspectral imaging with data analysis using artificial intelligence (AI) and has recently developed a bench-top system that predicted the concentrations of four groups of mycotoxins in whole, or ground, samples of grains. The cooperator is attempting to improve their technology by developing a hand-held version of their device. In this agreement the Agency PIs will work with the cooperator to develop models to predict aflatoxins in corn using a hand-held hyperspectral imaging device. Development of the models will require many samples that range from non-contaminated to highly contaminated with aflatoxins. The Ageny PIs will inoculate corn with strains of Aspergillus flavus to produce infested kernels. Whole kernel corn samples will be imaged using the handheld device, ground, and imaged a second time. The ground samples will be shipped to a third-party analytical laboratory for determination of aflatoxin content using a widely accepted reference method (high-performance liquid chromatography-fluorescence). The images will be uploaded to the cooperator's system for processing, and along with the reference values of aflatoxin content, used to develop models to predict aflatoxin content in corn. Because some of the aflatoxins have a native fluorescence, additional experiments will be conducted to determine if the AI system can directly measure, rather than predict, the toxin content. Appropriate material transfer agreements and data transfer agreements will be utilized.