【摘要】 利用傅立叶光谱仪的智能光纤传感器 ,探讨了近红外光无损检测水果糖度的方法。通过主成分回归、偏最小二乘法和逐步回归法三种多元校正算法对水果光谱数据的分析 ,得出PLS模式水果糖度预测值和真实值的相关系数为 0 86 ,标准校正误差为 0 14 ,标准预测误差为 0 97,PCA模式水果糖度预测值和真实值相关系数为 0 85 ,标准校正误差为 0 4 7,标准预测误差为 1 2 7,而SMLR模式的水果糖度预测值和真实值相关系数为 0 72 ,标准校正误差为 0 72 ,标准预测误差为 2 10 ,研究结果表明 ,在 12 5 0 0～ 380 0cm- 1 光谱范围内应用光纤传感器技术进行苹果糖度无损检测具有可行性更多还原

【Abstract】 Near infrared spectroscopy method with optical fiber sensor as a non destructive measurement for Fuji apples sugar content was evaluated. The objective of this research was to investigate a fiber sensing technique in interactance mode for rapid acquisition of spectral information to predict the interior quality of fruit. Three multivariate calibration techniques including partial least squares analysis(PLS)?principal component analysis(PCA) and stepwise multiple linear regression(SMLR) were used for spectral data analysis. Using PLS models, the correlation coefficient was 0 86, the standard error of calibration was 0 14 and the standard error of prediction was 0 97. Using PCA models, the correlation coefficient was 0 902, the standard error of calibration was 0 47 and the standard error of prediction was 1 27. Using SMLR models, the correlation coefficient was 0 72, the standard error of calibration was 0 72 and the standard error of prediction was 2 10. It is suggested that fiber optic sensor technique be feasible to nondestructively determinate apple sugar content in the range of 12 500～3 800 cm -1 .更多还原