Shortwave infrared hyperspectral imaging for the determination of pork adulteration in minced beef and lamb

Abstract

Meat adulteration raises religious concerns and economic, quality, and safety issues. Therefore, ensuring that no unnecessary meat is added to the labeled primary meat is important. This study used shortwave hyperspectral imaging (SWIR HSI) working at 895–2504 nm to predict the level of pork adulteration in minced beef and lamb samples. Minced beef and lamb meat samples were adulterated with 2%–50% (w/w) minced pork. Unsupervised principal component analysis (PCA) was used to analyze and visualize the spectral data to determine their similarities. Spectral data were used to develop a partial least squares regression (PLSR) model to predict pork adulteration in minced beef and lamb. PLSR, with several spectral preprocessing methods, resulted in R2 0.97 and RMSE 2.47–2.55 for mixed pork and beef (Pork + Beef) model and R2 0.84–0.91 and RMSE 5.07–6.10 for mixed pork and lamb (Pork + Lamb) model. The developed PLSR model was then applied to each pixel in an image to obtain chemical maps to visualize the presence of pork in minced beef and lamb. This study demonstrated the potency of SWIR HSI to be applied in the sorting machine in meat industries to detect adulteration in meat.