摘要
采用Daubechies正交小波滤波和样条卷积分峰联用方法对信噪比S/N≥ 2的含噪重叠双峰进行滤噪和分峰处理。滤波部分采用阈值法 ,较详细地探讨了阈值的选取规则 ;分峰部分利用样条卷积法 ,并给出了其算法公式。处理速度较快 ,算法也较简单。通过对模拟和实验信号的处理 ,发现该联用方法很有效 ,处理后的峰的峰位置误差≤ 3 0 % ,峰面积误差≤ 3 0 %
A coupled method of Daubechies Orthogonal Wavelet Filtration and Spline Convolution Resolution (SCR) is presented to process overlapped peaks with higher noises. The rules of selecting filtration threshold are discussed. And the algorithmic formula of SCR is given. Simulating signals of S/N(signal to noise ratio)≥2 are processed and the results meet the analytical request: The relative errors of peak intensity and peak area are less than 3 0% respectively. Besides the data from capillary electrophoresis are processed and the peaks with higher noises are separate successfully. The coupled method was programmed in MATLAB and its algorithm is simple. The sentences are less than 50 rows and the speed of processing data is also fast.
出处
《计算机与应用化学》
CAS
CSCD
北大核心
2001年第5期460-462,共3页
Computers and Applied Chemistry
基金
国家自然科学基金 ( 2 9975 0 33)
广东省自然科学基金 ( 980 340 )
关键词
正交小波滤波
样条卷积分峰
含噪重叠峰
分辨
分析化学
滤噪
信号处理
daubechies orthogonal wavelet filtration
spline convolution resolution
overlapped peaks with higher noises
