【作者】 黄明月； 吴海云； 靳皓； 董桂梅； 杨延荣； 杨仁杰；

【Author】 HUANG Ming-yue;WU Hai-yun;JIN Hao;DONG Gui-mei;YANG Yan-rong;YANG Ren-jie;College of Engineering and Technology,Tianjin Agricultural University;

【机构】 天津农学院工程技术学院；

【摘要】 将近红外透射和漫反射光谱数据进行融合,并结合偏最小二乘法建立了掺杂牛奶的判别模型。分别准备蒙牛、三元和完达山纯牛奶样品各40个、三个品牌掺杂尿素牛奶样品各40个。室温下,采集所有样品的近红外透射光谱和漫反射光谱。在此基础上,分别建立了三个品牌掺杂牛奶的透射、漫反射和透射-漫反射融合偏最小二乘判别模型,三个模型对未知样品的判别正确率分别为85.9%,91%和93.6%。研究结果表明:相比于单一的近红外透射和漫反射光谱,融合近红外透射-漫反射光谱所建立的模型可以减弱不同品牌牛奶对模型预测能力所造成的干扰,能提供更好的判别结果。更多还原

【Abstract】 In this paper,the fusion of near-infrared transmission and diffuse reflectance spectra combined with the partial least square is used to establish the discrimination model of adulterated milk.Forty mengniu milk samples,40 Sanyuan milk samples,40 Wandas-an milk samples,and 40 three brands milk adulterated with urea,were prepared respectively.Near-infrared transmission spectra and diffuse reflectance spectra of all samples were collected at room temperature.On this basis.the three discriminant models were established using transmission,diffuse reflection,and the fusion of transmission-diffuse reflection spectra,respectively.The discriminant accuracy of the three models for unknown samples was 85.9%,91% and 93.6% respectively.The results show that the fusion of transmission-diffuse reflection spectra,which can reduce the interference caused by different brands milk,can provide better discrimination results.更多还原