说明：  为了研究适合近红外光谱无损检测脐橙VC 含量的光谱预处理方法, 比较了11 种光谱预处理方法对偏最小二乘法( PLS) 建模预测效果的影响。结果表明, 小波消噪效果最好, 模型预测值与标准值的相关系数R 达到0. 957 4, 交叉验证预测均方差仅为3. 90, 最优光谱波段为7 501. 7～5 449. 8 cm- 1, 主成分数为8。小波消噪是脐橙VC 含量近红外光谱无损检测的有效光谱预处理方法。
(In order to search an appro priate pret reatment method to nondest ruct ive measure vitamin C co ntent of umbilical orange, the ef fect of 11 pretr eatment methods to the model based on PLS has been compar ed. T he result s show ed that the best ef fect w as wav elet do-noised. It s R w as 0. 957 4, it s R MS ECV w as 3. 90, it s best w av e band w as 7 501. 7～5 449. 8 cm- 1, and it s principal element w as 8. T he w avelet do-noised was an ef ficient method to nondest ruct ive measure vitamin C content of umbilical o rang e w ith near-inf rared spect roscopy.)