摘要
采用连续样条小波变换确定重叠信号的组分峰的数目和峰位置,将样条小波最小二乘法加以改进,使其成为既能滤出噪声又能分峰的新方法。对信噪比为1的重叠峰进行处理,得到满意结果。该方法只需对信号进行一次处理,处理速度快,且数据点无须为2n(n∈Z)。应用于毛细管电泳电导信号的处理,得到满意结果。
A novel algorithm of modified spline wavelet least square(SWLS)was proposed in this paper and applied to process noisy overlapped signals.After the number and positions of overlapped peaks were continuˉously spline wavelet transformed(CSWT),the noisy overlapped peaks could be separated(S/N=1)directly and the noises were removed at the same time.Compared with the overlapped signals,the noises were filtered and the resolution of the signals processed by SWLS was enhanced.The signals only need to be processed only once.This method is simple and quick.The number of data points does not have to be2 n (n∈Z).The reˉsults of resolving the capillary electrophoresis overlapped peaks with heavy noises were satisfactory.
出处
《分析测试学报》
CAS
CSCD
北大核心
2005年第3期86-89,94,共5页
Journal of Instrumental Analysis
基金
华南理工大学自然科学基金资助项目(E512670)
