| 徐洪庆,赵凌云,宋晓冉,宋晓丹,吕媛媛,孙 辉.基于色谱相似性分析模型的水质紫外可见光谱检测方法[J].,2024,(21):4041-4046 |
| 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 of Water Quality by UV-VIS Spectrum Based on Chromatographic Similarity Analysis Model |
| 投稿时间：2024-04-20 修订日期：2024-05-14 |
| DOI：10.13241/j.cnki.pmb.2024.21.007 |
| 中文关键词: 色谱相似性 分析模型 水质检测 紫外可见光谱 Q统计量 |
| 英文关键词: Chromatographic similarity Analysis model Water quality testing Uv-visible spectrum Q statistics |
| 基金项目: |
|
| 摘要点击次数: 72 |
| 全文下载次数: 26 |
| 中文摘要: |
| 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。 |
| 英文摘要: |
| ABSTRACT 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. Conclusion: 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. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
