RAPID ASSESSMENT OF ITALIAN POLYFLORAL HONEY QUALITY BASED ON NON-DESTRUCTIVE SPECTRAL SENSORS

One of the main products of beekeeping is honey, a natural sweet substance produced by bees (Apis mellifera) from the nectar of plants or from the secretions of living parts of plants or from substances secreted by sucking insects found on the living parts of plants. The aim of this work was to assess whether the use of non-destructive spectral sensors, such as a VIS-based spectrophotometer and NIR-based devices, can assess different chemical, sensory, and instrumental colour quality traits of the polyfloral honey samples from three main ecological areas of Italy: SL = South Lowland, below 600 meters above sea level (asl); NM = North Mountain, above 600 meters above sea level; NL = North Lowland, below 600 m above sea level. Moreover, the NIR data were used to build a PLS-DA model to assess a proposed overall merit quality class and the quality class of four sensorial traits. The experimental design of my master's thesis (MSc degree) examined 215 samples of multifloral honey collected from various Italian beekeepers. The geographical and botanical origin, as well as the sensory test, were carried out by trained experts from the Italian "National Honey Observatory" at the end of the 2022 harvest season, during the "Three Drops of Gold - Great Honeys of Italy" competition. The sensory evaluation was conducted according to an official and certified protocol, and related data were kindly permitted to be used for this MSc dissertation. After a slight thermal treatment (at 40 °C for approximately 30 minutes) of the honey samples, the spectral data were collected using a portable VIS spectrophotometer (operating in the 450-700 nm range) and a benchtop NIR device (operating in the 850-2500 nm range). The chemical traits were estimated using predictive equations from NIR spectroscopy performed in previous studies of UNIPD researchers. The chemical and organoleptic variables have been used to define a quality index useful for assessing the overall quality of honey as a quick tool to distinguish samples collected in different ecological areas. Furthermore, the degree of correlation between the instrumental chromatic coordinates and the sensory characteristics was assessed. Finally, a PLS-DA model was made via NIR spectroscopy data aimed at discriminating the honey overall merit quality class and the quality class of the four sensory traits.