In-Line Microwave Nondestructive Evaluation of Packaged Food Products via the Support Vector Machine Algorithm

This paper presents a novel approach based on electromagnetic waves (EM) to classify food packages that hold water as one of the main ingredients from the inside into contaminated or uncontaminated products. A non-destructive technique that can handle a real-time food production line is proposed to achieve this goal. This technique combines the operation of a microwave sensing system (MW) with a machine learning (ML) classifier. An accuracy of 100% has been obtained from training the aforementioned ML tool on a dataset constructed from the retrieved scattering parameters of about 500 measuring samples.