为了比较4 ℃、15 ℃两种贮藏温度下滩羊肉pH值的变化, 优选出滩羊肉贮藏期间的最优模型,采用pH酸度计测量样本pH值, 建立两种贮藏温度下传统动力学的零级和一级模型; 应用近红外（900~1 700 nm)高光谱成像采集两种贮藏温度下滩羊肉的光谱数据, 剔除异常值后进行光谱预处理; 使用连续投影算法（SPA)提取特征波长, 建立全波段和特征波长的偏最小二乘（PLSR)预测模型; 对比分析得到的最优光谱模型与动力学模型相结合, 确定滩羊肉光谱动力学模型。结果表明,4 ℃和15 ℃的传统动力学模型的相关系数分别为0.502和0.912; 4 ℃下原始光谱经PLSR建模后效果最优, 相关系数Rc为0.821, Rp为0.863, 15 ℃经SG-S(3,7)+De-trending（4)预处理后经PLSR建模效果最优, 相关系数Rc为0.876, Rp为0.819。因此, 高光谱结合传统动力学的模型检测羊肉pH值的方法是可行的, 该模型可以预测15 ℃下滩羊肉的贮藏期。

In order to compare the change of the pH value of mutton at 4 ℃ and 15 ℃ storage temperature, the optimal model of mutton during the storage period was optimized. The pH values of the samples were measured by pH acidity meter, the zero-order and first-order models of the traditional kinetics were established under two storage temperatures. The spectral data of mutton at two storage temperatures were acquired by near-infrared hyperspectral imaging(900-1 700 nm), and the spectral pretreatment was carried out after removing outlier. The feature wavelength is extracted by successive projection algorithm(SPA), and the partial least squares(PLSR) prediction model of full wavelength and feature wavelength is established. The optimal spectral model was combined with the kinetic model, and the spectral kinetic model of mutton was determined. The results show that the correlation coefficients of the conventional kinetic models at 4 ℃ and 15 ℃ were 0.502 and 0.912, respectively. The original spectrum was the best after modeling by PLSR at 4 ℃, and the correlation coefficient Rc was 0.821, Rp was 0.863. After the SG-S(3,7)+De-trending(4) pretreatment, the model was optimized by PLSR at 15 ℃. The correlation coefficient Rc was 0.876 and Rp was 0.819. It is feasible to detect the pH value of mutton by hyperspectral and traditional kinetic model, and the model can predict the storage period of mutton at 15 ℃.