小波变换用于近红外光谱性质分析

以汽油研究法辛烷值分析为例，研究了小波变换在近红外光谱分析中的应用。对近红外光谱的小波特性、小波变换参数以及变量提取方法进行了详细研究。研究结果表明：光谱噪音、有用信息和背景分别分布在小波高、中和低频区域；母小波函数对性质分析结果影响很大；小波变换可以同时扣除光谱背景、去除噪音和压缩变量，具有运算速度快、分析精度高以及无需去噪后处理等优点，在近红外光谱分析中具有很好的应用前景。

Application of Wavelet Transform on the Near Infrared Analysis

This paper is directed to study application of wavelet transform (WT) on the near infrared (NIR) analysis. The wavelet properties of NIR spectra, the determination of WT parameters and the methods extracting useful variables were studied in detail. Results show that backgrounds, useful signal and noise are located in low-, middle- and high-frequency region respectively. The mother wavelet function has great effects on the performance of WT. WT can correct background, conduct denoising and compress the variables; consequently it can increase speed of quantity analysis and improve the analysis precision, showing great prospects in NIR analysis.