CAJ | 학술논문

以蛋白核小球藻( Chlorella pyrenoidosa)作为鉴别载体，利用共聚焦显微拉曼光谱仪分别获取生长在除草剂草甘膦、丁草胺污染水体以及正常水体的蛋白核小球藻β-胡萝卜素的拉曼光谱信息，对2种除草剂进行了鉴别.利用预处理后的光谱信号，建立偏最小二乘回归( PLS)预测模型及线性判别分析( LDA)分类模型.当阈值为±0.3时，全波段建立的PLS模型对草甘膦和丁草胺的预测正确率高达83.33%，特征峰建立的LDA分类模型对2种除草剂的分类正确率均达到了100%.结果表明，利用蛋白核小球藻为载体对丁草胺和草甘膦2种除草剂进行鉴别是可行的，且LDA分类模型更适合除草剂的分类研究.
이단백핵소구조( Chlorella pyrenoidosa)작위감별재체，이용공취초현미랍만광보의분별획취생장재제초제초감련、정초알오염수체이급정상수체적단백핵소구조β-호라복소적랍만광보신식，대2충제초제진행료감별.이용예처리후적광보신호，건립편최소이승회귀( PLS)예측모형급선성판별분석( LDA)분류모형.당역치위±0.3시，전파단건립적PLS모형대초감련화정초알적예측정학솔고체83.33%，특정봉건립적LDA분류모형대2충제초제적분류정학솔균체도료100%.결과표명，이용단백핵소구조위재체대정초알화초감련2충제초제진행감별시가행적，차LDA분류모형경괄합제초제적분류연구.
This study investigate the method to identify two herbicides using Chlorella pyrenoidosa as a carrier and obtained the Raman spectral information of β-carotene of Chlorella pyrenoidosa grown in glyphosate-polluted water, butachlor-polluted water and normal water, respectively. The partial least-squares ( PLS ) model and linear discriminant analysis( LDA) classification model were established by the spectral signal after preprocessing. When the threshold is ± 0. 3, the correct prediction rate to glyhosate and butachlor reached 83. 33% by using the PLS model based on the whole band, and correct classification rate of the two herbicides can reach 100% by using LDA classification model based on characteristic peaks. The results show that using Chlorella pyrenoidosa as a carrier to identify heribicides glyphosate and butachlor is feasible and LDA classifi-cation model is more suitable for classification of the two herbicides.