近红外光谱结合主成分分析鉴别不同产地的南丰蜜桔

摘要

采用近红外光谱结合主成分分析（PCA）建立不同产地南丰蜜桔鉴别模型，实现不同产地南丰蜜桔的快速鉴别。分别研究一个果园内不同位置的蜜桔，洽湾、市山和白舍等南丰县三个不同乡镇的南丰蜜桔，福建邵武、广西柳城和江西南丰等三个不同省份的南丰蜜桔之间的差异，蜜桔保存时间对主成分分析模型的影响。结果表明同一个果园内不同位置的蜜桔不存在明显差别，不同产地的蜜桔有很好的分类效果，蜜桔的短时间保存对近红外光谱的主成分分析模型不会产生明显影响。不同的光谱预处理方法对主成分分析模型产生较大影响，多元散射校正（MSC）结合二阶导预处理得到的主成分分析投影具有最佳的分类效果。该研究可为南丰蜜桔的产地鉴别提供一种参考方法。%The identification of Nanfeng mandarins from different origins was developed by using near infrared spectroscopy (NIR) and principal components analysis (PCA) .Mandarins from different places in one orchard ,and from different orchards in three towns of Nanfeng county were studied .Also differences among Shaowu ,Liucheng and Nanfeng were investigated .Fur-thermore ,the effect of storage time of mandarins on the PCA model was considered .The results demonstrate that there was no clear diversity of the mandarins in one origin but great differences existed among different ones .And the storage time of manda-rins played little role in the discrimination model .The method of multiple scattering correction (MSC) coupled with second de-rivative was selected to build PCA discrimination models compared with other data pretreatment methods .The proposed model would be a reference method for origin identification of Nanfeng mandarin .