论文部分内容阅读
为了建立苹果质地参数的近红外光谱无损检测方法,研究了不同建模方法、不同导数处理结合不同散射及标准化处理对苹果质地参数(果皮强度、果皮脆性和果肉平均硬度)近红外漫反射光谱无损检测模型性能的影响。结果表明,在全光谱范围内,采用改进偏最小二乘法、一阶导数和标准归一化和趋势变化法散射处理相结合的光谱预处理方法建立的果皮强度、果皮脆性和果肉平均硬度校正模型效果较好,果皮强度、果皮脆性和果肉平均硬度的预测相关系数分别为0.8577、0.8880、0.8885,预测标准误差分别为0.5132、0.9060、0.1827。因此,通过近红外漫反射光谱对苹果质地参数的快速无损检测具有可行性。
In order to establish NDT NDT detection method of apple texture parameters, different modeling methods, different derivation treatments and different scattering and normalization treatments were applied to evaluate the texture parameters of apple (pericarp strength, pericarp brittleness and average hardness of pulp) near-infrared diffuse reflectance spectroscopy Detect the impact of model performance. The results showed that the peel intensity, pericarp brittleness and the average hardness correction model established by the improved partial least square method, the first derivative and the standard normalization and the trend change method were used to establish the spectral model of pericarp in the whole spectrum. The results showed that the correlation coefficients of pericarp strength, pericarp brittleness and average hardness of pulp were 0.8577, 0.8880 and 0.8885, respectively. The predicted standard errors were 0.5132, 0.9060 and 0.1827 respectively. Therefore, the rapid non-destructive testing of apple texture parameters by near-infrared diffuse reflectance spectroscopy is feasible.