Wavelet-based Identification Of Overlapping Chromatographic Peaks And Related Optimization

With the food and environmental safety issues increasingly severe, people have an increasing demand on food, medicines and contaminants detection, the development of high-performance chromatography analytical instruments without delay. The signal processing is an integral part of chromatography analytical, and its task is to effectively filter out the noise from the signal analysis spectrum, enhanced the anti-interference ability of analytical instrumentation system, suppress noise and detect weak signals. Since the actual chemical measurement system is generally a multi-component system of coexistence, interference between the components is inevitable, when multiple peak position is close to the test substance, the overlap peak phenomenon will be inevitably. The existence of overlapping peaks directly affect the chromatographic signal spectrum can be qualitative and quantitative analysis, therefore the decomposition of overlapping peaks is usually the key to the analysis of chromatographic data. After years of development and improvement, the overlapping peaks identification methods become more sophisticated due to the introduction of a variety of advanced algorithms. However, with the increased complexity of the test substance, the degree of the overlap in the chromatogram is increasing, the requirements of the processing algorithms performance on the high degree overlaps peaks separation is also increasing. The method based on wavelet transform and quadratic differential is proposed in this paper, because the traditional algorithms’resolving effect of high degree of overlapping peaks can not meet the actual demand. To improve performance of the algorithm, this paper studies the characteristics of wavelet function and then proposes the correlation coefficient and energy leakage as the rule of the best wavelet selection. The result show that the method proposed in this paper can effectively improve the overlapped peaks resolving effect. In addition, this paper also designed the signal acquisition module and use the VS2010programming software to design the corresponding supporting software system, and the preliminary experiments verify that it basically can be applied in practical engineering.