Scanning Foods
by Judy Purdy

The money lost on bad onions is enough to bring tears to your eyes.
Because a single bad onion — or apple — can ruin an entire bag or bin-full, the ripple effect costs processors and consumers hundreds of thousands of dollars each year.

That’s why E. William Tollner is adapting CAT scan medical technology to fresh produce processing. His produce scanner, now in the early stages of development, can peer inside fresh fruits and vegetables, finding bad apples and onions long before defects are visible to the naked eye.

“Scanning technology would enable industry to put maximum quality products in the hands of consumers,” said Tollner, a UGA biological and agricultural engineering professor.

It also could become an important safeguard given the increasing amount of foods the United States imports from other parts of the world. “As we look into the future, we will have less and less control over the production of many of our foods because they come from other countries,” Tollner said.

As the food industry continues to mechanize, processors are looking for ways to minimize bruises and to identify defects. Removing a bruised or otherwise defective onion or apple can help preserve the quality and shelf life of the rest of the bunch. And scanning technology also can help processors determine optimal uses for rejected products — whether they end up in jams, sauces or frozen juices.

“It’s not like all the ‘bad’ produce is going to the compost pile,” Tollner said. “It’s just not saleable as fresh. A bruised or lightly diseased onion could still be trimmed, dehydrated and used in a condiment.”

Tollner is among several UGA scientists using digital images from a CAT scanner to identify food quality problems. The team also includes Yen-Con Hung, an expert in the physical properties of foods, and Stan Prussia, a post-harvest handling expert, as well as several graduate students. Their funding has come from a variety of state, federal and private sources, including the U.S. Department of Agriculture and the Georgia Onion Committee.

The UGA scanning research started with Red Delicious apples, because they are prone to water core — a condition of high internal moisture.

“Water core makes the apple sweet and tasty but also hard to store,” said Tollner, whose scanning research began a dozen years ago.

Scanning for water core could help processors know which apples to sell and which to store. But to ensure the scanning technique was effective, Tollner’s team first had to compare thousands of digital images to the actual damaged produce.

“We cut a lot of apples and onions open and looked until we became confident that what we were seeing in the scanner image was what we were seeing on the fruit itself,” Tollner said. “With apples, we were detecting the water core conditions and old bruises. With onions, we were looking for bruises and internal rot.”

Then the researchers linked computers, math models and software to the scanner technology. Using known defects, the scientists trained the computer network to recognize the defects and found the scanning process was 90 percent accurate.

The next step will be to send a signal to a device that automatically removes bad fruit from the processing stream.

The scientists have begun work with a private company to develop a prototype they hope to market within the next five years. Using a scanner the size of two old-fashioned file cabinets, they project their prototype could scan and cull one to two fruits per second.

“A person can’t do all that this machine will do,” Tollner said. “One person can see a defect on the screen this fast, but can’t reject the fruit this fast.”

The scanners use a very low-level X-ray that Tollner said poses no risk in food processing.

“I see no reason for food safety concern,” he said. “This fruit has been growing out in the environment all summer long, being exposed to common background radiation. Our X-ray technique is about equal to cumulative background radiation. We use minimal intensity and very quick exposure that is less than a dental X-ray.

“Consumers are beginning to balance the whole question of irradiated food against food safety,” he said. “Many processed foods are already X-rayed for foreign objects.”

Tollner and Hung also are conducting research to validate measurements, made by UGA graduate student Jimmy Wong, of how a hamburger patty cooks. Wong built a small vertical fry pan that could be installed in the CAT scan machine.

“We can scan the hamburger patty as it cooks and note the changes,” Tollner said. “[Wong] is mapping properties of the image to thermal conductivity, which will enable a model to be made of the patty while it cooks.”

UGA researchers are finding other potential uses for the technology. Tollner and Suchendra Bhandarkar, a UGA computer science professor, are experimenting with scanners for lumber mills. A rapid, three-dimensional image could determine the value of a log before a saw blade ever touches it, compensating for defects, minimizing waste and maximizing lumber cut to sizes in greatest demand.

“I predict it’s going to be more important to be able to do quality checks that will enable us to put a good product out there for the consumer,” Tollner said.

For more information, access
http://www.engr.uga.edu/research/groups/ndsensing.