基于色谱相似性分析模型的水质紫外可见光谱检测方法  

Detection Method ofWater Quality by UV-VIS Spectrum Based on Chromatographic Similarity Analysis Model

作　　者：徐洪庆 赵凌云 宋晓冉 宋晓丹[3] 吕媛媛 孙辉 XU Hong-qing;ZHAO Ling-yun;SONG Xiao-ran;SONG Xiao-dan;LüYuan-yuan;SUN Hui(Qingdao Branch Center,Operation and Maintenance Center of Shandong Water Transfer Project,Qingdao,Shandong,266000,China;Hydrological center of Qingdao,Qingdao,Shandong,266000,China;Water conservancy vocational college of Shandong,Rizhao,Shandong,276800,China;School of Environmental and Municipal Engineering,Qingdao University of Technology,Qingdao,Shandong,266520,China)

机构地区：[1]山东省调水工程运行维护中心青岛分中心,山东青岛266000 [2]青岛市水文中心,山东青岛266000 [3]山东水利职业学院,山东日照276800 [4]青岛理工大学环境与市政工程学院,山东青岛266520

出　　处：《现代生物医学进展》2024年第21期4041-4046,共6页Progress in Modern Biomedicine

摘　　要：目的:研究基于色谱相似性分析模型的水质紫外可见光谱检测方法,为准确检测出污染水质的各种物质提供依据。方法:数据分析方法包括利用小波变换与主成分分析法处理紫外可见光谱仪采集的水质紫外可见光谱,依据Q统计量阈值计算部分光谱观测点的离群点,通过改进二分K均值聚类算法,依据色谱相似性分析模型计算多元紫外可见色谱相似度实现水质污染物聚类。结果:经过处理能够获取去噪和降维后的光谱数据,判断水质异常,为离群点赋予权重获得概率值,运算异常概率,检测出水体样本中是否存在相关类型的污染物。结论:该方法处理后的光谱更加简洁,准确检测出水质异常,且检测出水体样本中是否存在不同类型的污染物,具有较强应用性。Objective:To investigate the UV-VIS spectrum detection method for water quality based on chromatographic similarity analysis model,in order to accurately detect various substances that pollute water quality.Methods:The data analysis methods included using wavelet transform and principal component analysis to process the water quality UV-VIS spectrum collected by UV-VIS spectrometer,calculating the outliers of some spectral observation points according to the threshold of Q statistics,and calculating the multivariate UV-VIS chromatographic similarity according to the chromatographic similarity analysis model by improving the bipartite K-means clustering algorithm.Results:After processing,spectral data after denoising and dimensionality reduction could be obtained to determine abnormal water quality,assign weight to outliers to obtain probability values,calculate abnormal probability,and detect whether relevant types of pollutants exist in water samples.Conclusions:The spectrum processed by this method is more concise,accurately detecting water quality anomalies,and detecting the presence of different types of pollutants in water samples,which has strong applicability.

关 键 词：色谱相似性 分析模型 水质检测 紫外可见光谱 Q统计量 

分 类 号：X824[环境科学与工程—环境工程]