中国激光, 2020, 47 (11): 1111001, 网络出版: 2020-11-02   

利用已知混合物拉曼光谱改善混合物成分识别精度的方法 下载: 914次

Method for Improving Identification Accuracy of Components in Mixtures Using Raman Spectra of Known Mixtures
作者单位
1 江南大学轻工过程先进控制教育部重点实验室, 江苏 无锡 214122
2 北京卓立汉光仪器有限公司, 北京 101102
摘要
构建已知纯净物光谱数据库,计算待识别混合物光谱与数据库中各纯净物光谱的相似度,是利用拉曼光谱技术进行混合物成分分析的常用策略。受测量仪器性能及待测混合物所含成分相互干扰的影响,待识别混合物中所含物质成分的光谱与数据库中对应的纯净物光谱相比会有不同程度的失真,从而给基于纯净物光谱数据库的组分鉴别带来极大困难。针对这一问题,提出了一种使用已知混合物光谱数据来改善混合物成分识别精度的方法。首先利用纯净物拉曼光谱谱峰的位移和半峰全宽信息,将已知混合物的光谱谱峰与其所含有的具体物质对应;然后基于谱峰拉曼位移、半峰全宽和谱峰强度分别构建纯净物、已知混合物和待识别混合物的特征参数,并利用模糊隶属度函数计算待识别混合物光谱与纯净物光谱、已知混合物所含物质光谱的相似度;最终根据光谱相似度确定待识别混合物中含有的疑似组分。基于204种纯净物和8种已知混合物光谱数据库,对81种混合物进行了识别,结果表明:所提方法可降低由光谱失真导致的相似度计算误差,提高识别准确率;相比于纯净物数据库搜索策略,本文方法的识别精度由76.34%提高到了92.83%。
Abstract
A common strategy for mixture composition analysis based on Raman spectroscopy is to construct the spectral database of pure substances and calculate the spectral similarity of the mixture to be identified and the pure substances in the database. However, influenced of the performance of the measuring instrument and the mutual interference of the components of the mixture, the spectrum of the substance contained in the mixture to be identified will have different degrees of distortion compared with the corresponding pure substance spectrum in the database, bringing great difficulties in component identification. To address this problem, a method to improve the identification accuracy of components in mixtures using the spectral data of known mixtures is proposed herein. The spectral peak information, including Raman shift and full width at half maximum, of the pure substance in the database is used to study the correspondence of the spectral peaks of the known mixture to the specific substances of the mixture. The spectral feature parameters of the pure substance, the known mixture, and the mixture to be identified are constructed using the Raman shift of the spectral peak, the full width at half maximum, and the peak intensity, respectively and the fuzzy membership function is used to calculate the spectral similarity between the mixture to be identified, the pure substance, and the substances contained in known mixture based on calculated feature parameters. Furthermore, suspected components contained in mixture to be identified are determined based on the spectral similarity. Based on the spectral database of 204 pure substances and 8 known mixtures, the experimental results for 81 unknown mixtures reveal that the proposed method can reduce the calculation error of similarity caused by spectral distortion and can improve the identification accuracy. Compared with search strategy based on the database of pure substance, the identification accuracy obtained using the proposed method increases from 76.34% to 92.83%.

季明强, 朱启兵, 黄敏, 张丽文, 雷泽民, 张恒. 利用已知混合物拉曼光谱改善混合物成分识别精度的方法[J]. 中国激光, 2020, 47(11): 1111001. Ji Mingqiang, Zhu Qibing, Huang Min, Zhang Liwen, Lei Zemin, Zhang Heng. Method for Improving Identification Accuracy of Components in Mixtures Using Raman Spectra of Known Mixtures[J]. Chinese Journal of Lasers, 2020, 47(11): 1111001.

本文已被 3 篇论文引用
被引统计数据来源于中国光学期刊网
引用该论文: TXT   |   EndNote

相关论文

加载中...

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!