近红外光谱技术快速测定鹅肉新鲜度

目的:应用近红外光谱技术快速检测鹅肉的新鲜度,评价指标包括总挥发性盐基氮和p H值。方法:采集完整冷鲜鹅肉的近红外光谱(950~1 650 nm),光谱经多种校正预处理后,采用偏最小二乘法建立鹅肉新鲜度的定量预测数学模型。结果:对于这2种指标均采用标准常态变量结合偏最小二乘法所建立模型的预测效果最好,总挥发性盐基氮和p H值定量校正数学模型的模型决定系数分别为0.727、0.991,内部交互验证均方根误差分别为3.666、0.028。用此模型对预测集20个样品进行预测,预测值与实测值的相关系数分别达到0.976、0.705,预测值平均偏差分别为-0.240、-0.024,预测值和实测值之间没有显著性差异(P0.05)。结论:近红外光谱作为一种无损快速的检测方法,可用于评价鹅肉新鲜度。

Rapid Determination of Goose Meat Freshness Using Near Infrared Spectroscopy

Objective: To determine goose meat freshness based on total volatile base nitrogen (TVB-N) and pH by near
infrared (NIR) spectroscopy. Methods: Near infrared spectra (950–1 650 nm) of goose meat were collected, and then
sequentially subjected to multiple correction pretreatment, multiple linear regression, and principal component regression for
the establishment of quantitative prediction mathematical models for evaluating goose meat freshness based on TVB-N and
pH by partial least squares regression. Results: The models obtained by standard normal variate (SNV) combined with partial
least squares regression exhibited the best prediction performance with a coefficient of determination for calibration of 0.727
and 0.991, and a root mean square error of cross validation (RMSECV) of 3.666 and 0.028 for TVB-N and pH, respectively.
The correlation coefficients between predicted and measured values of TVB-N and pH for 20 samples were 0.976 and 0.705,
and the average deviations were ?0.240 and ?0.024, respectively, suggesting no significant difference (P > 0.05) between
predicted and measured values. Conclusion: NIR spectroscopy as a rapid nondestructive detection method can be used in the
evaluation of goose meat freshness.