采用紫外可见(ultraviolet/visible, UV/Vis)光谱技术对水体中有机物浓度的指标化学需氧量(chemical oxygen demand, COD)进行快速检测, 将收集到的135份水样进行UV/VIS波段全光谱扫描, 应用Savitzky-Golay (SG)平滑算法, 经验模态分解算法(empirical modedecomposition, EMD)和小波分析(wavelet transform, WT)对提取出的光谱数据进行去除噪声处理, 为了简化模型, PLSR建模得到的6个潜在变量(LVs)作为偏最小二乘支持向量机(LS-SVM)的输入建立COD预测模型, LS-SVM模型的预测集决定系数r2为0.82, 预测均方根误差RMSEP为14.82 mg·L-1。 说明使用LVs作为LS-SVM建模输入, 可以准确快速检测水产养殖水体中的COD含量, 为将来实现水产养殖水质COD含量的在线检测以及其他水质参数的快速测定奠定了基础。

Ultraviolet/visible (UV/Vis) spectroscopy was studied for the rapid determination of chemical oxygen demand (COD), which was an indicator to measure the concentration of organic matter in aquaculture water. In order to reduce the influence of the absolute noises of the spectra, the extracted 135 absorbance spectra were preprocessed by Savitzky-Golay smoothing (SG), EMD, and wavelet transform (WT) methods. The preprocessed spectra were then used to select latent variables (LVs) by partial least squares (PLS) methods. Partial least squares (PLS) was used to build models with the full spectra, and back-propagation neural network (BPNN) and least square support vector machine (LS-SVM) were applied to build models with the selected LVs. The overall results showed that BPNN and LS-SVM models performed better than PLS models, and the LS-SVM models with LVs based on WT preprocessed spectra obtained the best results with the determination coefficient (r2) and RMSE being 0.83 and 14.78 mg·L-1 for calibration set, and 0.82 and 14.82 mg·L-1 for the prediction set respectively. The method showed the best performance in LS-SVM model. The results indicated that it was feasible to use UV/Vis with LVs which were obtained by PLS method, combined with LS-SVM calibration could be applied to the rapid and accurate determination of COD in aquaculture water. Moreover, this study laid the foundation for further implementation of online analysis of aquaculture water and rapid determination of other water quality parameters.