Novelty Sensor using Integrated Fluorescence and Dielectric Spectroscopy to Improve Food Quality Identification

Sensors are essential components of any modern device, and currently, any modern device has many types of sensors for several applications. However, there is a trade-off between cost and accuracy in sensors, and this frontier should be overcome with the use of new sensor types and algorithms for data processing. This study presents a novelty sensor by combining fluorescence and microwave dielectric spectroscopy into the same physical structure to improve food quality identification. The proposed sensor incorporates a planar circular complementary split-ring resonator for the dielectric measurement and an optical path to fluorescence spectroscopy into the same physical structure to make measurements simultaneously. The proposed sensor was tested to identify adulterated honey. Using the data from fluorescence spectroscopy and microwave dielectric spectroscopy, a greater diversity of data from the electromagnetic spectrum made it possible to improve the accuracy of honey adulteration identification while using fewer samples to train machine learning classification algorithms. The proposed sensor was tested with a low-cost Vis-NIR sensor; it achieved an accuracy of 100% using 20 training samples only with the Bilayered neural network classification algorithm. Therefore, the proposed sensor paves the path for material characterization and identification by combining several spectroscopy techniques in the same structure.