中文摘要：

菜籽品种是决定蔬菜品质的主要因素之一,为了快速鉴别菜籽品种,提出一种菜籽品种的可见近红外光谱鉴别方法。在450～1 000 nm光谱范围内,采集了6种菜籽的可见近红外光谱。采用主成分载荷曲线法,从全光谱变量中筛选出17个特征光谱变量,作为线性判别分析（LDA）的输入向量,建立了菜籽品种LDA鉴别模型,且模型判别结果优于主成分判别结果,误判率为2.42%。结果表明,采用可见近红外光谱技术结合线性判别分析方法,进行菜籽品种的快速鉴别是可行的。

英文摘要：

The seed varieties of vegetable are one of the major factors that decide the quality of the vegetables. A discrimination method of vegetables＇ seeds by visible and near-infrared spectroscopy （Visible- NIR） was proposed for rapid discriminating the varieties of vegetables＇ seeds. The spectra of six vegetables＇ seeds were collected by reflection mode in the range of 400-1 000 nm. Seventeen variables were chosen by curve peak and valley trends of principal components loadings. And linear discrimination analysis （LDA） models were developed with the input vectors of seventeen variables. The discriminating results of LDA models were superior to the principal component analysis （PCA） with the error recognition rate of 2.42%. The results suggest it is feasible for rapid discrimination varieties of vegetables＇ seeds by visible-NIR tech- nology and LDA method.