Neural Network and Microwave Sensing for Food Contamination Monitoring

In this paper, a technique covering the lacks in currently adopted food safety devices is presented, through a microwave-sensing prototype combined with a neural network. An antennas array, composed by 6 low-cost printed circuit boards, surrounds the product to analyze. Their positions and number have been chosen as a trade-off between an optimal coverage of the volume of interest and the physical constraints of an industrial device, i.e., the conveyor belt, moving at production speed. The signals are recorded and used to train a neural network, resulting in an overall 99.45% success rate in classification.