LAUSR

Abstract Pork subcutaneous fat (PSF) is used for the development of several products, and the evaluation of the qualitative and quantitative characteristics currently involves lengthy procedures. In this study, the discriminant and predictive capability of near-infrared hyperspectral imaging were tested on PSFs, which were collected from the belly, ham, loin, and shoulder. The spectral information from the hyperspectral images was used to discriminate amongst these fats as well as predict their iodine values (degree of fat unsaturation) using partial least squares discriminant analysis (PLS-DA) and partial least square regression (PLSR). The result showed that PLS-DA was outstanding in discriminating PSFs with classification error in the validation dataset ≤ 0.05%. For the IV, result for the IV measurement showed a good accuracy (R2p = 0.77 and RMSEP = 1.83-gram I2/100g fat). This study showed that the hyperspectral imaging technique has the potential to discriminate PSFs from different locations of the pork carcass and also evaluate the IV.