Non-destructive spectroscopy-based technologies for meat and meat product discrimination - A review

Abstract

Consumers' confidence in products of animal origin is highly subjected to the quality guarantees offered by the manufacturing and retail industries. Traditionally, meat quality evaluation has been conducted through destructive, time-consuming and chemical-dependent protocols. Smart methodologies based on the non-destructiveness and/or non-contact with the samples, such as spectroscopy-based technologies, arise as an alternative promising tool. This comprehensive overview includes literature published in the last decade applying spectroscopy-based techniques in the Visible (Vis) and near-infrared (NIR) regions of the spectrum (Vis-NIR), either individually or combined with imaging (hyperspectral imaging, HSI), to classify meat and meat products based on ante- or postmortem factors. First, a brief introduction to the fundamentals of Vis-NIRS and HSI is included. Secondly, the main applications of Vis-NIRS and HSI technologies for meat qualitative purposes only are discussed. The Vis-NIRS and HSI have been successfully used in lab scale studies (> 90 % overall accuracy) to discriminate meat and meat products according to antemortem (feeding system, species, origin and breed) and postmortem (freshness, meat quality, label claims) factors. Recently, spectral data collected with handheld Vis-NIR equipment have become more frequent, although the use of portable HSI has not been widely explored. From the studies reviewed, the main concern regarding spectral data is to shorten modelling handling times, including strategies to both extract optimal wavelengths from NIR and compress spectral data from HSI. Despite the efforts made to overcome instrumentation and data processing challenges, a gap remains to be covered up to a real-time implementation in industrial line quality control.

Keywords: Classification; Hyperspectral imaging; Machine-learning techniques; Meat; Meat products; Visible-near-infrared spectroscopy.