Gas chromatography-ion mobility spectrometric classification of vegetable oils based on digital image processing

In this paper, a headspace instrument equipped with gas chromatography-ion mobility spectrometry (GC-IMS) was used to classify three kinds of vegetable oils in cooperation with chemometric tools. The procedure contained direct loading of the vegetable oil sample into a vial, headspace generation, and automatic injection of volatile organic components into GC-IMS device. A total of 187 oil samples were detected by GC-IMS, and Otsu's threshold segmentation and colorized difference methods were adopted to realize automatic peak detection of two-dimensional matrix and comparative visualization for further chemometric pretreatment. Based on the obtained data, principal components analysis showed that 95.77% of sample information could be explained by the first two principal components. Moreover, the oil samples were divided into calibration set (n = 130) and prediction set (n = 57), and the model built by the k-nearest neighbors algorithm showed that the recognition accuracy of calibration set was 100% and the recognition accuracy of prediction set was 98.24%. These results verify that digital image processing methods applied to GC-IMS datasets could preserve chemical information and support qualitative analysis. Thus, GC-IMS technique can be considered a vanguard and reliable tool for recognition of different types of common vegetable oils.