光谱学与光谱分析, 2015, 35 (3): 622, 网络出版: 2015-05-21
花生种子品质的可见-近红外光谱分析
Quality Analysis of Peanut Seed by Visible/Near-Infrared Spectra
可见-近红外光谱 花生种子 小波分析 主成分分析 线性判别分析 鉴别 Visible/Near-infrared Spectroscopy Peanut seed Wavelet analysis Principal component analysis(PCA) Linear discriminant analysis Discrimination
摘要
利用600~1 100 nm波段研究花生品种的可见-近红外反射光谱, 对选取的三种具有代表性的花生种子进行实验。 使用近红外光纤光谱仪采集光谱数据, 对原始光谱进行小波分析以提取光谱特征, 再用主成分分析方法进行聚类分析, 最后把每一个样品的前4个主成分得分作为识别模型的输入, 品种类别作为模型的输出, 以马氏距离作为判别函数, 建立了线性判别分析模型。 对于每个品种的50个样品, 随机挑选30个样本作为训练集, 剩余的20个样本作为预测集。 该识别模型对3个花生品种的平均正确识别率为95%。 表明该方法能有效的识别花生种子, 得到较好的分类效果, 为花生种子品质的区分和鉴别提供了一种新方法。
Abstract
In this paper, three representative varieties of peanut seeds were selected for the experiment based on visible/near-infrared reflectance spectroscopy living in the wavelength rang from 600 to 1 100 nm.Firstly, spectral datas ware collected by the near-infrared fiber optic spectrometer, and the spectral features of the original spectral dates were extracted by the wavelet analysis.Then the principal component analysis (PCA) was used for cluster analysis of spectral features. Finally, the four principal components were applied as the inputs, the varieties category as the output and the Mahalanobis distance as the discriminant function of the recognitionmodel, so a linear discriminant analysis model was established.In the 50 samples of each varieties, 30 samples were randomly selected as the training set, and the remaining 20 samples as the predictor set. The recognitionmodel for three peanut varieties have a recognition rate of 95% on average. As the experimental results show that this method is reliable and effectively, and a new method to distinguish and discriminate the quality of peanut seeds was put forword.
郑田甜, 孙腾飞, 曹增辉, 张骏. 花生种子品质的可见-近红外光谱分析[J]. 光谱学与光谱分析, 2015, 35(3): 622. ZHENG Tian-tian, SUN Teng-fei, CAO Zeng-hui, ZHANG Jun. Quality Analysis of Peanut Seed by Visible/Near-Infrared Spectra[J]. Spectroscopy and Spectral Analysis, 2015, 35(3): 622.