Review team checks progress of aflagoggle project

By Allison Floyd
University of Georgia, Peanut & Mycotoxin Innovation Lab

What if bad maize kernels or groundnuts glowed so that a processor could simply pick them out?

Soon, technology might make that possible.

“We want to create a rapid, low-cost detection method that may be used in developing countries as a means for aflatoxin detection,” said Haibo Yao, a Peanut & Mycotoxin Innovation Lab scientist and associate research professor for Mississippi State University.

One of the challenges with detecting aflatoxin in crops – a nasty byproduct of common mold in the soil – is that mold doesn’t spread evenly. A single nut or kernel may have a high level of contamination, while all those around it are toxin-free. And while highly contaminated nuts or kernels may show clear signs, the human eye can’t always see the damage mold left behind.

A portable, fluorescence spectral-based technology might “see” the bad kernels or nuts and allow a machine or person to pick them out.

At first, Yao envisioned goggles that would allow a tester to see the fluorescence emitted by the toxin. That concept has evolved into a handheld device, currently a small box, that uses UV LEDs to cause the contaminated kernels to fluoresce and the camera in a tablet to detect and display the image.

“Our method is non-invasive. It only looks at the surface to gauge the UV-fluorescence rate phenomena in the maize kernel,” Yao said.

In the experiment, 5 kg of maize was divided into 100 samples, each weighing about 50 g. Each sample was mixed with one or more contaminated maize kernels. The mixed samples were imaged with a fluorescence hyperspectral imaging system, and a sorting algorithm was applied to separate the samples into contaminated and clean groups.

Then, the clean maize was tested.

By locating and removing just 1 percent of the 5 kg samples – the bad kernels – the researchers were able to reduce the overall aflatoxin contamination by 30 percent.

An independent review team recently evaluated the progress and made suggestions for the next steps of the project, including using the device to screen peanut samples.

Published March 3, 2016