关键词: 注水肉糜；低场核磁共振；磁共振成像；主成分分析；预测模型

Abstract: In this paper, water-injected ground meat was detected quickly and non-destructively by using low-field nuclear magnetic resonance (LF-NMR) and magnetic resonance imaging (MRI) techniques. Based on their relaxation characteristics, qualitative and quantitative analyses were carried out on ground meat with different proportions of water injected. The results showed that with the increase in the proportion of water injection, the single-component relaxation time Tw value of water-injected ground meat increased linearly (R2 = 0.953 5), and the multi-component relaxation time (T22 and T23) increased significantly (P < 0.05). LF-NMR combined with principal component analysis (PCA) revealed that ground meat samples injected with different proportions of water were distinctly distinguished on the principal component score plot. The partial least squares regression (PLSR) and principal component regression (PCR) models showed determination coefficient R2 for validation set of 0.92, and root mean square error (RMSE) of 1.27% and 1.25%, respectively. The proton density-weighted images of ground meat indicated that as the proportion of water injection increased, the color of the pseudo-color map gradually changed from blue to yellow and red, thereby allowing to visually distinguish the water-injected ground meat. All above results indicate that LF-NMR and MRI techniques can be used to quickly and accurately determine whether or not ground meat is water-injected.

Key words: water-injected ground meat; low-field nuclear magnetic resonance; magnetic resonance imaging; principal component analysis; prediction model