将小波变换应用于重叠化学信号的分辨与处理,由于在噪声较大的情况下二阶导数法很难从噪声中区分出微弱信号峰,故通过小波卷积法确定各个组成峰的位置.在信噪比相对较低的情况下明显优于二阶导数法,然后利用小波滤波器处理数据,处理后的峰可达到基线分离,峰位置和面积基本不变.将此方法应用于单一成分的正丁酮、正丁醛以及乙醇、正丙醇、异丙醇、正丁醇、正戊醇和仲辛醇6种醇类物质的混合物处理,得到了满意结果.

Wavelet transformation was applied to resolve overlapped chemical signals. By the use of wavelet convolution,the positions of overlapped signal peaks were accurately resolved and determined,even when the S/N ratio was relatively low,while it was not achieved by the method of 2nd derivation. In the further treatment with wavelet filtering,the peaks were clearly separated from the baseline with no change in peak positions and peak areas. Both the algorithms of wavelet convolution and wavelet filtering were applied to the testing of alcohol mixture containing ethanol,2-propanol,1-propanol,1-butanol,1-pentanol and 2-octanol,mixture of butanone and butanal,as well as the single-componental butanone and butanal. As proved by experimental result,algorithm of wavelet transformation was an effective way to resolve the overlapped signals and making the determination very satisfactory.