基于支持向量机回归的3种常见有毒赤潮藻荧光识别技术

发布时间：2019-06-06 11:04

【摘要】：本文基于色素萃取液三维荧光光谱(EEM),利用支持向量机技术建立了3种常见有毒赤潮藻的荧光识别分析技术。选取3种有毒浮游藻在不同光照条件下进行培养,获得其色素萃取液的三维荧光光谱(EEM);利用Daubechies7(db7)小波对浮游藻色素萃取液EEM进行分解,并利用Bayesian判别分析选择最佳识别特征光谱;利用支持向量机回归建立有毒赤潮藻的荧光识别分析技术。该技术对单种藻的识别正确率为100%,对于88个测试集混合样品,当赤潮藻相对含量为10%、20%、30%、40%、60%、80%和90%时,海洋卡盾藻的识别正确率分别为66.7%、66.7%、100.0%、100.0%、100.0%、100.0%和100.0%,东海原甲藻的识别正确率分别为50%、100%、100%、100%、100%、100%和100%,球形棕囊藻的识别正确率分别为0.0%、50.0%、25.0%、60.0%、100.0%、100.0%和100.0%。可见,当3种有毒赤潮藻相对含量处于较低水平,也即处于赤潮发展期时,该技术就可对其快速进行准确的识别测定。
[Abstract]:In this paper, based on the three-dimensional fluorescence spectrum (EEM) of the pigment extract, the fluorescence identification and analysis technology of three kinds of common toxic red tide algae was established by using the support vector machine technology. Three kinds of poisonous planktonic algae were cultured under different light conditions to obtain the three-dimensional fluorescence spectrum (EEM) of the pigment extract, and by using the Daubechies7 (db7) wavelet, the phytoplankton pigment extract EEM was decomposed, and the best identification characteristic spectrum was selected by the Bayesian discriminant analysis. The fluorescence identification and analysis technology of the poisonous red tide algae was established by the regression of the support vector machine. The identification accuracy of the technology to the single algae is 100%, and for 88 test sets mixed samples, when the relative content of the red tide algae is 10%,20%,30%,40%,60%,80% and 90%, the identification accuracy of the marine cabococcus is 66.7%, 66.7%, 100.0%, 100.0%, 100.0%, respectively, The accuracy of the identification of the original dinoflagellates in the East China Sea was 50%,100%,100%,100%,100%,100% and 100%, respectively. The accuracy of the identification of the spherical-brown algae was 0.0%, 50.0%, 25.0%, 60.0%, 100.0%, 100.0% and 100.0%, respectively. It can be seen that when the relative content of three kinds of toxic red tide algae is at a lower level, that is, in the red tide development period, the technology can quickly and accurately identify and determine the red tide.
【作者单位】：中国海洋大学海洋化学理论与工程技术教育部重点实验室;
【基金】：国家自然科学基金项目(41376106) 国家科技重大专项项目(2012ZX07501) 山东省自然科学基金项目(ZR2013DM017)资助~~
【分类号】：X55;X835

【相似文献】