基于小波变换的重叠色谱峰检测方法研究  

Novel method for detecting overlapping chromatographic peaks using wavelet transform

作　　者：熊智新[1] 胡上序[1] 路文初[2] 

机构地区：[1]浙江大学化学工程学系,浙江杭州310027 [2]浙江大学分析测试中心,浙江杭州310027

出　　处：《浙江大学学报（工学版）》2004年第8期1072-1077,共6页Journal of Zhejiang University：Engineering Science

摘　　要：为了构建一种色谱数据处理方法,可以同时解决包含肩峰的复杂重叠色谱峰群检测中的噪声滤除、基线校正和谱峰辨识等问题,作者利用小波变换多分辨分析及奇异性检测原理,提出了一种色谱信号处理的新方法.通过对仿真和实验数据处理的结果表明:该方法检测重叠色谱信号无须专门的滤波和基线校正步骤,可以确定出色谱图中包含肩峰在内的复杂峰群各特征点位置和峰形情况.用该方法开发的检测算法只需设置两个参数,方便了计算机的自动处理.This study aimed at finding a unified method for denoising,correcting baseline driftage and resolving complex overlapping chromatographic peak group with overlapped and shoulder peaks.Singularity (detection) based on multiresolution analysis of wavelet transform was developed to process chromatography signals.A series of novel algorithms for peak recognition were proposed and successfully applied to the simulated and practical signals of complexly overlapping chromatographic peaks. Shoulder peaks can be (identified) by the characteristic points on the chromatogram. Besides, there are no requirement of individual denoising and baseline correction procedures for executing the detection, and only two parameters are (needed) to implement the operation,which makes the algorithm convenient for computers to execute automatically. These features are superior to those of the conventional detection method based on derivatives.

关 键 词：小波变换 色谱峰检测 多分辨分析 奇异性 

分 类 号：O652.9[理学—分析化学] TP274.23[理学—化学]