摘要
提出一种利用近红外光谱技术进行鸡蛋种类快速、无损鉴别的新方法。选用7500—4000cm-1的光谱,采用标准正态变量变换(SNV)后作主成分分析(PCA),选取前10个主成分作为模型输入,种类类别作为模型输出,分别建立了3种鸡蛋种类的线性判别法(LDA)和支持向量机(SVM)鉴别模型,所建模型均能较好的对鸡蛋种类进行鉴别,SVM模型效果优于LDA模型,其预测集正确识别率达97.44%。结果表明,近红外光谱技术可用于鸡蛋种类的快速、无损鉴别。
A rapid non-destructive method for kind discrimination of egg by near infrared spectroscopy（NIR） was developed.The raw spectra of 7500—4000cm^-1 were pre-processed by standard normal variable transformation（SNV） and principal component analysis（PCA）,then a support vector machines（SVM） model and a linear discriminate analysis（LDA） of 3 kind of egg model were built with the first 10 PCA as inputs and the kind of eggs as outputs.The kind of eggs can be well discriminated by both two models,and the SVM model was better than LDA with the correct recognition rate in prediction sets of 97.44%.The results show that NIR is rapidly and non-destructively applied to kind discrimination of eggs.
出处
《光谱实验室》
CAS
CSCD
2012年第5期2699-2702,共4页
Chinese Journal of Spectroscopy Laboratory
基金
中国计量学院-浙江苍南仪表厂仪器科学研究生教育创新示范基地(YJ2008023)
