【摘要】 近红外光谱常被用于含氢有机物质等的物化参数测量,可以提供丰富的结构和组成信息在复杂溶液的光谱定量分析中更是被广泛应用。然而在人体血液等复杂溶液的近红外光谱分析中,强大的背景信息造成的噪声干扰和冗余变量的存在,严重影响着样品的光谱测量和分析,影响着分析的效率和准确度。如何消除背景噪声等的干扰来提高分析准确度已经引起高度重视,近几年来,国内外学者提出了许多基于化学计量学方法的相关方法。从光谱预处理、变量优化和建模分析三方面,以传统的化学计量学方法出发,总结和分析这些方法在人体血液等复杂溶液的近红外光谱定量分析的应用和各自的特点,为提高光谱定量分析准确度的研究提供参考。更多还原

【Abstract】 Near-infrared spectroscopy has rich structure and composition information and is often used to measure hydrogen-containing organic substances’ physical and chemical parameters. In recent years, it has been widely used in the quantitative analysis of complex solutions. However, the near-infrared spectroscopy analysis of complex solutions such as human blood, noise interference and redundant variables caused by strong background information seriously affect the spectral measurement and analysis of the sample itself, and affect the efficiency and accuracy of the analysis. Therefore, eliminating the interference of background noise to improve the accuracy of analysis has attracted great attention. In recent years, scholars at home and abroad have proposed many related methods based on chemometric methods. This article focuses on spectral preprocessing, variable optimization and modeling analysis. On the one hand, starting from traditional chemometric methods, we summarize and analyze the application and respective characteristics of these methods in the near-infrared spectroscopy quantitative analysis of complex solutions such as human blood, and provide a reference for research on improving the accuracy of quantitative spectral analysis.更多还原