- UF researchers are leveraging artificial intelligence to tackle the global problem of food waste.
- The scientists analyzed how chemical changes in aging broccoli correspond with visual data picked up by a highly sensitive camera called a hyperspectral camera.
- The researchers hope to use these insights to develop a device that can scan produce for freshness.
When you’re picking out fruits and vegetables at the grocery store, you probably try to choose what looks freshest.
But looks can be deceiving. Produce that looks great on the outside can sometimes be far less fresh than you might think — and might end up in your trash, unused.
That’s why researchers at the University of Florida are using artificial intelligence to identify early signs of decay that aren’t visible to the naked eye. One day, they hope to create a device that packing houses and grocery stores can use to scan produce for freshness and ultimately reduce food waste.
In a new study recently published in journal Plant Phenomics, a team led by Tie Liu and Alina Zare takes the first steps toward developing such a device. Leveraging artificial intelligence techniques, the researchers showed how chemical changes in aging broccoli correspond with subtle changes in the vegetable’s appearance.
“Broccoli is high in compounds called glucosinolates, which have antioxidant properties and are one reason why broccoli is so good for you. In this study, we found that after broccoli is harvested, the levels of glucosinolates fluctuate in a consistent way as the plant ages,” explained Liu, an assistant professor in the UF/IFAS horticultural sciences department.
The scientists then compared those chemical patterns to changes detected by a hyperspectral camera.
“A normal camera picks up reflectance of red, green and blue wavelengths, but hyperspectral cameras measure a whole spectrum of light, including wavelengths we can’t see, such as infrared. All that additional data can help us uncover patterns that would have been invisible before,” said Zare, a professor in the UF Herbert Wertheim College of Engineering.
“We then use AI and machine learning to look for patterns in the hyperspectral data that sync up with the chemical fluctuations we are interested in within the broccoli over time,” said Zare, who leads the Machine Learning and Sensing lab in the department of electrical and computer engineering.
“Down the line, we want to use these insights to develop a device that can scan for the visual signatures we know are associated with freshness,” Liu added.
While the datasets involved this experiment were not large enough to require UF’s AI supercomputer, HiPerGator 3.0, future studies may call for that level of enhanced computing power, the researchers said.
Moving forward, Zare and Liu hope to use their combined expertise to analyze the aging processes of other produce such as avocados and potatoes. The overall research project, called FreshID, is funded by a nearly half-million dollar, four-year grant from the U.S. Department of Agriculture’s National Institute of Food and Agriculture (NIFA).
“About 40% of food produced in the U.S. is wasted. A significant proportion of that waste and loss occurs between the time the produce is harvested and when it gets to the consumer. The FreshID project aims to help make sure that food gets to the consumer at the right time,” Liu said when the project kicked off last year.