摘要
【目的】对不同品种薰衣草进行快速鉴别,为其正确使用提供可靠的科学方法。【方法】采用傅利叶红外光谱法测定93个属4种不同品种薰衣草花样品的红外谱图,以1600 cm^(-1)～1665 cm^(-1)范围内的吸收峰为指标,应用主成分分析(PCA)和径向基神经网络(RBFNN)进行数据分析,对不同品种薰衣草进行鉴别。【结果】对不同品种薰衣草的傅利叶红外光谱进行数据分析,主成分分析(PCA)结果表明,前3个主成分的累积可信度已达93.43%,可将薰衣草分为4个品种,基于FTIR谱的主成分分析能够在一定程度表征出薰衣草在不同品种的多样性分化,在对薰衣草品种进行主成分分析的基础上,选用64个薰衣草花样本建立径向基神经网络模型(RBFNN),余下29个作为预测样本,所建模型的拟合率和预测品种的识别率均为100%。【结论】实验表明主成分分析(PCA)对不同品种薰衣草具有较好的聚类作用,径向基神经网络模型(RBFNN)能对薰衣草进行较好的识别,说明该方法能快速无损的鉴别薰衣草,为薰衣草的品种识别提供了一定的科学依据。
[Objective] The aim of this study was to provide a fast and reliable scientific approach to identified different varieties of Lavender. [Method] Fourier transform infrared （FTIR） spectroscopy was used for obtaining vibrational spectrum of 93 flower samples which belong to four different Lavender species. Based on the indices of wave number absorbance from 1600 cm^-1 to 1665 cm^-1. Different varieties Lavender were analyzed by Fourier transformation infrared spectroscopy （FTIR） combined with the principal component analysis （PCA） and RBF neural network. [Result] Lavender samples were clustered into 4 classes by the methods of principal component analysis （PCA）. FTIR spectrum of principal component analysis to some degree reflect Lavender in different species diversity. Based on the principal component analysis of Lavender, the RBF neural network model be established with 64 Lavender samples, the remaining 29 as a predictor of the sample, the model fitting rate and prediction of species recognition rate is 100%.[Conclusion] Lavanderwere analyzed by principal component analysis （PCA）, which could intuitively distinguish Lavender varieties. RBF neural network model provided good recognition for the varieties identification of Lavender. The method can be used to fast identify varieties of lavender and provided some scientific basis.
出处
《计算机与应用化学》
CAS
CSCD
北大核心
2013年第6期655-658,共4页
Computers and Applied Chemistry
基金
新疆维吾尔自治区自然科学基金(2012211A016)
