基于近红外光谱与SIMCA和PLS-DA的水稻品种鉴别

Identification of Rice Varieties Using NIR Spectroscopy and SIMCA, PLS-DA Methods

以垦粳5号、垦粳6号、垦粳9号和鸿育001-1共4个水稻品种为研究对象,采用近红外光谱技术分别结合SIMCA和偏最小二乘法判别分析法(PLS-DA)对4个水稻品种进行鉴别。采用SIMCA分类法,实现了4个水稻品种100%的区分;采用PLS-DA方法对校正样本建立判别模型,其校正集结果和参考值之间的相关系数最小值为0.95,验证集结果与参考值之间的相关系数最小值为0.94,对验证集中4个水稻品种的识别率为100%。试验结果表明,应用近红外光谱技术结合SIMCA分类法和PLS-DA法均可实现对水稻品种的快速鉴别。

Four varieties of rice, including Kengjing 5, Kenjing 6, Kenjing 9 and Hongyu 001-1, seeds were selected as research object in this paper. Two discrimination methods, such as Soft Independent Modeling of Class Analogy （SIMCA） and Partial Least Squares-Discriminant Analysis （PLS-DA）, were employed to identify the four varieties of rice seeds based on the spectroscopy data of the rice seeds. The discrimination result of four kinds of rice by using SIMCA method is 100%. The correlation coefficients were more than 0.95 and 0.94 for calibration and reference value, validation and reference value, respectively, and the recognition rate of four kinds of rice in the validation set was 100% when using PLS-DA discriminant method to establish discriminant model. The results indicated that the rapid identification of rice varieties can be achieved by using near infrared spectroscopy combined with SIMCA and PLS- DA.