一种基于EWMA-PCA的水质光谱数据标准化方法研究  

Application Research of Normalization Algorithm Based on EWMA-PCA in Standardization of Water Quality Spectral Data

作　　者：周思寒 胡新宇[1,2] 汤斌[1,2,3] 赵明富[1,2] 李奉笑[1,2] 汪仁杰 肖棋森 肖渝 ZHOU Si-han;HU Xin-yu;TANG Bin;ZHAO Ming-fu;LI Feng-xiao;WANG Ren-jie;XIAO Qi-sen;XIAO Yu(Key Laboratory of Modern Optoelectronic Detection Technology and Instrument,Chongqing University of Technology,Chongqing 400054,China;Chongqing Key Laboratory of Optical Fiber Sensing and Photoelectric Detection,Chongqing University of Technology,Chongqing 400054,China;Key Laboratory of Optoelectronic Technology and Systems of Ministry of Education,Chongqing University,Chongqing 400044,China)

机构地区：[1]重庆理工大学现代光电检测技术与仪器重点实验室,重庆400054 [2]重庆理工大学光纤传感与光电检测重庆市重点实验室,重庆400054 [3]重庆大学光电技术及系统教育部重点实验室,重庆400044

出　　处：《光谱学与光谱分析》2020年第11期3443-3450,共8页Spectroscopy and Spectral Analysis

基　　金：国家自然科学基金项目(61805029);重庆市社会事业与民生保障科技创新专项(cstc2017shmsA1497);重庆市基础与前沿研究计划基金项目(cstc2016jcyjA0308)资助。

摘　　要：模型传递对于解决由于样本与各仪器响应函数不同所导致的量测信号不一致具有重要意义,而解决模型传递的有效方法是仪器或数据标准化。针对现有的光谱标准化方法中,鲜有对紫外-可见吸收光谱的应用研究,且紫外-可见光谱法水质多参数检测光谱探测单元存在分辨率、精度、响应范围不统一,难以进行不同仪器间测试数据的比对及多参数数据拟合的问题,提出采用EWMA-PCA归一化算法,实现紫外-可见水质光谱在不同仪器上的模型传递。EWMA(exponentially weighted moving-average)是一种指数加权平均移动算法,用以寻找以较高概率产生观察紫外-可见水质光谱数据的系统发生树,最大概率复原理论紫外-可见水质光谱数据,使紫外-可见光谱特征不丢失、不偏移,减小由于数据处理对紫外-可见水质光谱数据的影响。采用不同浓度的邻苯二甲酸氢钾溶液,对日本滨松C10082CAH光谱仪、美国海洋光学Maya2000Pro光谱仪以及厦门奥谱天成ATP2000光谱仪进行对比测试实验。对比组1选取源机滨松C10082CAH光谱仪和目标机海洋Maya2000Pro光谱仪,对比组2选取源机滨松C10082CAH光谱仪和目标机奥谱天成ATP2000光谱仪,对比组3选取源机海洋Maya2000Pro光谱仪和目标机奥谱天成ATP2000光谱仪。三组比对实验结果表明,该算法能很好地应用于不同的比对光谱仪中,在采用EWMA-PCA归一化算法对水质吸收光谱数据标准化后,相关系数达到99.576 5%,方差达到0.082 3%,且波峰偏移量可降低至0.000 5%,基于EWMA-PCA归一化光谱标准化算法具适应性广、所需传递样本少、传递精度高等优点,研究结果对光谱法水质检测仪器的广泛应用具有重要理论指导意义和工程应用参考价值。Model transfer is important for solving the inconsistency of measurement signals due to differences in sample and instrument response functions,and an effective method for solving model transfer is the instrument or data standardization.For the existing spectral standardization methods,there are few applied research on UV-visible absorption spectrum,and the UV-visible spectroscopy water quality multi-parameter detection spectrum detection unit has inconsistent resolution,accuracy,and response range,which is difficult to perform between different instruments.The problem of the comparison of test data and the fitting of multi-parameter data,this paper proposes to use the EWMA-PCA normalization algorithm to achieve the model transfer of UV-visible water quality spectra on different instruments’(Exponentially Weighted Moving-Average)is an exponentially weighted average moving algorithm for finding phylogenetic trees with a high probability of generating UV-Visible water spectral data and recovering theoretical UV-visible water quality spectral data with maximum probability.The UV-Vis spectral characteristics are not lost or offset,reducing the impact of data processing on the UV-visible water spectral data.In this experiment,different concentrations of potassium hydrogen phthalate solution were used to compare and test the Japanese Hamamatsu C10082CAH spectrometer,the US Ocean Optics Maya2000Pro spectrometer and the Xiamen Optosky ATP2000 spectrometer.In the comparison group 1,the source machine Hamamatsu C10082CAH spectrometer and the target machine ocean Maya2000Pro spectrometer were selected,the comparison group 2 selected the source machine Hamamatsu C10082CAH spectrometer and the target machine Optosky ATP2000 spectrometer,and the comparison group 3 selected the source machine marine Maya2000Pro spectrometer and the target machine Optosky ATP2000 spectrometer.The experimental results of three sets of experiments show that the algorithm can be applied to different ratiometric spectrometers.After the EWMA-PCA no

关 键 词：紫外-可见水质光谱 指数加权平均移动-主成分分析 标准化 归一化 模型传递 

分 类 号：X52[环境科学与工程—环境工程]