Real-Time and Rapid Food Quality Monitoring Using Smart Sensory Films with Image Analysis and Machine Learning

Abstract

Detecting and reporting the quality of packaged food to the consumer in real-time can reduce the consumption of poor-quality food products. Current food quality detection and reporting technologies of perishable foods are usually expensive, complicated, and take a significantly long time to convey results. Herein, a real-time, simple, and user-friendly food freshness detection prototype was developed by combining a glycerol-based sensory film with unique visual color analysis and the k-nearest neighbors algorithm (KNN). We established the quantitative relationship between the pH, organic acid level, digital color variance, and food storage time. By measuring the color variations of sensor films as a function of food storage time, we demonstrate a technology to record the quantitative “RGB” values of sensory films to represent real-time and precise pH changes of the food sample and trace the real-time food spoilage degree (e.g., pork loin spoilage). Next, a quick-response (QR) reader with a center sensor film was designed to eliminate the environmental effects on the color variation in real time Furthermore, the KNN was implemented to classify food quality by training data from different sources. This study provides a technology well suited for large-scale food storage applications by combining a smart sensor film with a QR code design followed by image analysis and KNN. This real-time and rapid food quality monitoring technology will ultimately lead to a reduction in food waste and loss (FLW).