为了实现鲜枣常温贮藏期的硬度实时监测并对贮藏时间进行预测, 建立了室温下壶瓶枣贮藏期的近红外光谱硬度动力学模型。 基于二维相关光谱技术, 分析综合浓度影响下的壶瓶枣敏感波段, 优选的敏感波段为904, 980, 1 072, 1 200, 1 630, 1 941和2 215 nm。 分析不同贮藏天数的壶瓶枣果肉平均硬度, 并拟合出零级反应方程, 模型的相关系数为0.991 3, 标准误差为6.116×10-4。 鲜枣的贮藏过程中, 由于复杂的生理化学反应, 主要物质的含量发生变化, 并通过宏观的信息光谱特征和硬度得以体现。 将敏感波段下的光谱信息和贮藏期的硬度指标进行信息耦合, 建立壶瓶枣果肉硬度的偏最小二乘模型（partial least square, PLS）, 模型的预测精度RP为0.942 7, RMSEP为0.021 0。 进而以敏感波段的吸光度为自变量, 壶瓶枣果肉硬度指标为应变量, 进行多元回归定量分析, 建立近红外光谱硬度动力学模型, 模型的拟合优度即相关系数为0.983 9, 标准误差为0.024 9, 并在此基础上建立壶瓶枣贮藏时间与近红外光谱的线性回归关系。 研究表明, 基于二维相关光谱的硬度动力学模型可以实现对壶瓶枣果肉硬度指标的快速、 无损检测并实现其贮藏时间的预测。

In order to realize the real-time monitoring of hardness and predict the storage life of fresh jujubes in storage periods at room temperature, a kinetic model of hardness according to near-infrared (NIR) spectroscopy was established. The influence of integrated thickness on storage life of jujubes was explored using two dimensional correlation spectroscopy technology and the sensitive bands influenced by integrated thickness of 904, 980, 1 072, 1 200, 1 630, 1 941 and 2 215 nm respectively. The average hardness of the jujube pulp per day during storage was analyzed and then a zero level reaction equation was fitted, which was A=0.549 8-0.009 6A0-0.023 2t. The result showed that the correlation coefficient of the zero level reaction was 0.991 3 and the standard error was 6.116×10-4. The content of main material changed during storage because of the complex physiological and chemical reaction in the fresh jujube fruits. And this is represented through spectral characteristics and hardness. Coupled the spectral information in the sensitive band and the hardness index of the storage period, a partial least squares (PLS) model of jujube flesh hardness was established. The prediction precision of the model was 0.942 7, and the root mean squared error of prediction (RMSEP) was 0.021 0. And then, a kinetic model of jujube’ flesh hardness according to near-infrared (NIR) spectroscopy (A=0.549 8-0.009 6(0.793 6+0.655 1X1-3.804 2X2+2.372 2X3-1.884 2X4+3.637 3X5-1.041 7X6-1.327 8X7)-0.023 2t) was established through multivariate regression analysis, in which the spectral reflectances of characteristic bands are independent variables and the hardness indexes of jujube fruits are dependent variables. The correlation coefficient of this model was 0.983 9 and standard error was 0.024 9. The linear relation between storage life and near-infrared(NIR) spectroscopy was found to be t=23.698 2-0.413 8(0.793 6+0.655 1X1-3.804 2X2+2.372 2X3-1.884 2X4+3.637 3X5-1.041 7X6-1.327 8X7)-43.103 4(0.793 6+0.655 1X1t-3.804 2X2t+2.372 2X3t-1.884 2X4t+3.637 3X5t-1.041 7X6t-1.327 8X7t). The study shows that the near-infrared (NIR) spectroscopy technology that combining kinetic model of hardness can realize rapidly nondestructive testing of jujube’ flesh hardness index and the prediction of the storage time.