基于支持向量机对纸质食品包装材料的红外光谱研究
Identification Analysis of Paper Food Packaging by Infrared Spectroscopy Combined with Support Vector Machine
-
摘要: 为建立一种快速检验纸质食品包装材料的方法,对公安机关工作提供帮助,利用操作方便,可以重复使用并且无损材料的红外光谱技术对 76 个不同品牌的纸质食品包装材料样品进行无损检测.首先分析样品的红外光谱吸收峰的峰位,确定样品所含填料,进一步按照填料成分不同将样品进行人工分类;再通过K-MEANS法对样品聚类,验证了人工分类的准确性;最后根据聚类结果,构建支持向量机算法并建立纸质食品包装材料物证的鉴别模型.随机选择 50%的样品作为训练集,50%的样品作为测试集,结果表明,测试集的准确率为 97.368%,召回率为97.368%,精确率为 97.478%.实现了对包装纸的快速无损鉴别.此分类模型方法可操作性好,结果准确可靠,可为公安机关破案提供帮助.Abstract: To establish a rapid method for inspecting packaging paper and provide assistance to the work of public security organs,non-destructive testing was conducted on 76 different brands of packaging paper samples using infrared spectroscopy technology,which is easy to operate,reusable,and non-destructive.Firstly,the position of the infrared absorption peak of the sample is analyzed,and the filler contained in the sample is determined,and then the sample is manually classified according to the different filler composition,and then the samples are clustered by K-MEANS method to verify the accuracy of manual classification.Finally,based on the clustering results,a support vector machine algorithm is constructed and a identification model for packaging paper evidence is established.Randomly select 50%of the samples as the training set and 50%of the samples as the testing set.The results show that the accuracy of the testing set is 97.368%,the recall rate is 97.368%,and the accuracy rate is 97.478%.Realized rapid non-destructive identification of packaging paper.This classification model method has good operability,accurate and reliable results,and can provide assistance for public security organs in solving cases.