激光诱导击穿光谱技术对烟草快速分类研究

Different types of tobacco differ in element type and content. Based on Laser-Induced Breakdown Spectroscopy(LIBS), the atomic spectra were collected to realize fast classification and identification of different tobacco species combined with a support vector machine. In this paper, 9 kinds of cigarettes of different brands were selected and all characteristic peaks of spectral lines were extracted separately. Before classification, all feature peaks need to be preprocessed through multiple methods, including window smoothing to remove background and peak shift correction. After this, principal component analysis was used to reduce the dimension and the support vector machine method(SVM) was used to establish the classification model. Finally, classification results of 9 brands of cigarettes were obtained, with an average accuracy of 97.47%. It was proven that LIBS has great potential for applications in tobacco rapid identification, anti-counterfeiting identification and on-site rapid classification.