基于荧光法和多元回归分析的混合海藻识别方法研究

Hybrid Alga Identify Method Research Based on Fluorescence and Analysis of Multi-Element Regression Analysis

钱对单一海藻和混合海藻进行荧光激发光谱的二阶导数分析,发现它们的峰值点和谷点有很强的相似性。因此,用二阶导数荧光发射光谱的谷点和峰值点为特征点,取特征点对应的荧光发射光谱的相对强度的归一化数据为特征参数,进行混合海藻的线性多元回归分析,得到分配系数矩阵。去掉分配矩阵中的负分配系数和相对很小的分配系数,重新进行多元线性回归分析,直到达到目标模型为止,这样可以根据剩余的分配系数得到混合海藻中的单一海藻类目。

The two order differential coefficient for fluorescence inspired spectrum of single alga and hybrid alga was analyzed, and received the comparability of peak value point and coulee value point of them. Based on it, use two order differential coefficient for fluorescence inspired spectrum＇s peak value point and coulee value point as characteristics point, and use the normalized data of fluorescence inspired spectrum＇s relative intensity of characteristics point as characteristics coefficient, after linear multi-element regression analysis of hybrid alga, the distribute coefficient matrix was got. Then, delete negative and relative small coefficient, redo the linear multi-element regression analysis, until the target module was got, single alga class in the hybrid alga can be got according to residuary distribute coefficient.