GTRI

Case Study

Vision-based Screening System Detects Liner Material in Processed Foods

Published: November 3, 2007

Plastic liners and casings are used throughout the beef and poultry industries to ensure ingredients remain fresh and that meat does not come into contact with surfaces of cardboard or plastic containers that may harbor pathogens. Sometimes, despite extensive precautions, a part of a liner can tear off and become mixed in with a processed food. Liner pieces are particularly difficult to detect because they are often small, either transparent or nearly transparent, and coated with food product or the ingredient that was initially packaged in the container. Food processors have long sought a method to automatically detect these materials in processed food.

Researchers with Georgia Tech’s Food Processing Technology Division believe they have come up with a technique that could help resolve this challenge. With funding from Georgia’s Traditional Industries Program for Food Processing, the researchers have developed a prototype vision-based screening system that utilizes an FDA-approved additive to fluoresce the liner material when it is exposed to a particular wavelength of light. Specifically, the additive contains a UV marker that fluoresces when excited by a particular frequency of UV light. Some food components may also fluoresce in this band, but to date, the team has found this conflict to be minimal and in many cases to be resolvable through filtering.

The prototype system consists of a camera to detect the fluorescence from the marker and an illumination source that will cause the additive in the plastic material to fluoresce. The camera is equipped with a narrow band filter that is tuned to the fluorescent marker frequency and selected to reject the light from the main illumination source. The filter also rejects typical fluorescence from beef, pork, and poultry components under the illumination used to excite the additive.

The study team has conducted laboratory tests on the concept using ground beef. Three pieces of the liner surrogate that were approximately 5 mm square were tested, and were detected 20 out of 20 times. Detection was possible even when the liner was covered with fat from the meat. A smaller 2.5 mm square piece was also tested, but the detection rate for the smaller piece fell to about 75 percent. Liner material buried in meat, unfortunately, cannot be detected by the technique.

These initial tests indicate that occlusion is an important factor in letting pieces of liner pass through the system undetected. A side camera was added to the prototype cell to see if it could detect pieces the primary camera missed. The side camera was also fitted with a varifocal lens and filter. This second camera was placed perpendicular to the first camera and 40 cm from the centerline of the first camera. For large pieces of plastic material, the detection rate was again close to 100 percent with just the first camera. For smaller pieces that were about .5 cm2, the second camera improved the detection rate on average by 14 percent.

Researchers believe this development has strong potential for helping the food industry tackle a nagging challenge. “Keeping plastic liner materials and other plastic components out of food items is a serious issue for the food industry. This research demonstrates that high detection accuracies can be achieved using relatively affordable enhancement techniques that can be incorporated into the film manufacturing process,” says John Stewart, senior research engineer and project director.

“By combining this technique with specialized imaging technology, plastic liner and other marked plastics can be more readily detected on a processing line. Faster identification of plastic in a food stream will reduce the cost of each incident and allow producers to quickly respond to what is causing the contamination in the first place,” adds Stewart.