基于拉曼光谱的花粉检测及分类方法

Pollen Detection and Classification Method via Raman Spectroscopy

不同人群对不同种类的花粉存在不同的过敏反应,为此研究用于快速检测空气中花粉粒子及分类的方法。以常见花粉作为研究对象,利用拉曼光谱仪采集42种花粉样品的465条拉曼光谱数据,按照生物学分类划分为科间花粉及属间花粉并对其进行分类预测。将所得光谱数据预处理后,利用主成分分析提取光谱的特征信息,并建立支持向量机识别模型。对于科间花粉的预测结果准确率为97.75%,蔷薇科属间花粉的预测结果准确率为90.47%,说明拉曼光谱分析法对花粉分类鉴别具有可行性。

Different people exhibit different allergic reactions to various types of pollen.Therefore,this study presents a method to rapidly detect and classify the pollen particles in air.Herein,465 Raman spectroscopic data associated with 42 pollen samples were obtained using a Raman spectrometer by considering common pollen as the research object.Subsequently,they were categorized as interfamily and intergeneric pollen according to their biological classification,and then classified and predicted.After the obtained spectral data were preprocessed,principal component analysis was used for extracting the spectral characteristic information,and a support vector machine recognition model was established.The prediction accuracy of interfamily pollen is 97.75%,and the pollen prediction accuracy of the genus Rosaceae is 90.47%,which indicates that Raman spectroscopy can be used to classify and identify pollen in a feasible manner.