荧光光谱结合主成分分析对硅藻和甲藻的识别测定

Identification of Bacillariophyta and Pyrrophyta by Using Fluorescence Spectrum and Principal Component Analysis

将浮游植物三维荧光光谱与主成分分析方法相结合,针对我国东海常见的12种藻建立了硅藻和甲藻的识别测定技术。根据Fisher判别分析结果,选用三维荧光光谱的第一主成分得分谱作为硅藻和甲藻的识别特征谱,建立了浮游植物荧光特征谱库。在此基础上,利用非负最小二乘法进行识别测定,其中单门类样品共有78个,识别正确率为95%,回收率≥81%;2个门类的混合样品共有18个,识别正确率为83%,回收率≥80%。

In this paper, the identification techniques for 12 species of Bacillariophyta and Pyrrophyta often found in the East China Sea are developed by combining the three-dimensional fluorescence spectra of phytoplankton with principal component analysis method. Based on the results of Fisher linear discrimination analysis, the first principal component score spectrum of three-dimensional fluorescence spectrum is chosen as the identification characteristic spectrum for Bacillariophyta and Pyrrophyta, and the phytoplanktonic fluorescence characteristic spectrum bank is established. Some algal samples were identified by using nonnegative least square method, among them, 78 single division algal samples were tested with a correct identification rate of 95% and a recovery rate of no less than 81%; 18 mixed samples of two division algae were tested with a correct identification rate of 83 % and a recovery rate of no less than 80 %.