The Design And Implementation Of Spectral Peak Identification For Ion Mobility Spectrometry

In analytical chemistry field, we often need to analyze and process the obtained signal, including signal mass spectrometry, spectroscopy or gas chromatogram. The signal includes the starting point of the peak, end point of the peak, top of the peak, peak height, peak area, and other characteristics of these peaks. These point may confirm the various physical and chemical information of the tested substance and signal, such as physical nature and concentration, so the accuracy of peak detection means the veracity of tested material and signal.The Ion Mobility Spectrometry can be used for the fast qualitative and quantitative analysis of the species and content of the mixture in the mixture. It is widely used in the area of chemical warfare agents, drugs and explosives detection and environmental monitoring etc.. In this paper, we design and implement a peak identification system based on derivative method and wavelet transform and a qualitative identification system based graph theory for IMS.For the derivative method, firstly we use the mean filter method, the obtained spectral data is filtered, and then the first derivative and the second derivative of the spectral curve can be calculated. According to the change trend of the slope value, the starting point, the end point, the first inflection point and the top point can be identified. The signal are filtered using wavelet transform, which can filter out most of the noise through the choice of scale. Result in smoothing of the curve. Then, the algorithm uses the moving window method to detect the key points of wavelet transform curve. The algorithm identify the peak feature by using the key points and the logical correspondence between them in the original peak, and ultimately get the peak start of peaks, peak and the acromion inflection point, peak point, valley points, peak end and so on.The spectral peak set peak in the spectrum peaks clustering, K sub class, calculating each sub class related feature vector parameter information, save to spectral library, as the basis of recognizing and evaluate the toxic gases. By calculating the eigenvector of each sub class, and compare them with the spectral library of standard gas. If they are similar or even in perfectly matched within the scope, the qualitative identification of the material ion of FAIMS is realized.Through the analysis of experimental results we know that this system can find positioning and qualitative identification of the peaks accurately, so the system has achieved the expected goals.