Computer vision could help grill the perfect chicken

Advances in chemical sensor and computer vision technology could help restaurants automate the cooking process, reducing incidences of undercooked or contaminated chicken being served to consumers.

Doughman Headshot3 Headshot
looby | BigStock.com
looby | BigStock.com

Advances in chemical sensor and computer vision technology could help restaurants automate the cooking process, reducing incidences of undercooked or contaminated chicken being served to consumers.

Currently, chefs in commercial kitchens must rely on subjective judgement – what they see and smell – to detect when a chicken breast is properly cooked.

Automated quality control

The research team developed new technology that combines an electronic nose or e-nose with computer vision.

The research team designed a custom e-nose containing eight sensors able to detect smoke, alcohol, carbon monoxide and other compounds, and temperature and humidity. In previous studies, e-noses can sense if an apple and banana is rotten, and even tell the difference between types of cheese.

The team also created a custom computer algorithm that ‘looks’ at the chicken to decide when it was done cooking, based upon images of grilled chicken that were uploaded to the system.

“We employed pattern recognition techniques to derive classes depending on the cooking state train our system. Then we tested it using these images for detection of the cooking state,” Fedor Fedorov, a senior research scientist at Skoltech in Russia, explained.

“We believe this new technology can help deliver food that better meets consumer preferences using more objective methods.”

Results of the taste test comparison

A taste test compared the subjective ratings of graduate students to the results produced by the e-nose/computer vision combination for tenderness, juiciness, intensity of flavor, appearance and overall doneness.

The results revealed that the system was able to accurately identify the differences between undercooked, well-cooked and overcooked chicken breasts. The technology can probably be used on other parts of the chicken, like legs or wings, but the e-nose and computer vision algorithm would have to be retrained first, the researchers noted.

The research, supported by a Russian Science Foundation Grant, was published in the journal Food Chemistry.

Like what you just read? Sign up now for free to receive the Poultry Future Newsletter.

Page 1 of 180
Next Page