基于近红外光谱的胡椒产地鉴别方法研究
|
摘要
该研究旨在探索一种基于近红外光谱技术对胡椒产地进行分类的方法。收集海南、云南、广西、越南、马来西亚5个产地的胡椒共计300份样品,采集近红外光谱。采用小波去噪等方法对光谱进行预处理,通过支持向量机(support vector machine,SVM)、径向基神经网络(radical basic function,RBF)和线性判别分析(linear discriminant analysis,LDA)建立产地定性鉴别模型。研究表明,SVM和RBF神经网络模型鉴别准确率较好。db5小波预处理后仅选择7个主成分正确率达到100%的数据。结果表明基于近红外光谱的胡椒产地鉴别方法是可行的,预处理可以有效地提高近红外光谱胡椒产地鉴别模型的准确率。
This study aims to explore a method of classifying the places of origin of pepper based on near infrared spectrum.300 pepper samples from Hainan,Yunnan,Guangxi,Vietnam and Malaysia are collected and the near infrared spectrum of the samples are collected by NIR analyzer.Using wavelet denoising and other methods for spectra preprocessing,and then support vector machine(SVM),radical basic function(RBF)neural network and linear discriminant analysis(LDA)are applied to set up the identification models of places of origin of pepper.The accuracy of SVM model and RBF neural network model is great.The data after pretreatment with db5 wavelet selects only 7 principal components,with the classification accuracy is 100%.The results show that it is feasible to identify the places of origin of pepper based on near infrared spectroscopy,and the pretreatments could improve the accuracy of identification models of the places of origin of pepper effectively.
This study aims to explore a method of classifying the places of origin of pepper based on near infrared spectrum.300 pepper samples from Hainan,Yunnan,Guangxi,Vietnam and Malaysia are collected and the near infrared spectrum of the samples are collected by NIR analyzer.Using wavelet denoising and other methods for spectra preprocessing,and then support vector machine(SVM),radical basic function(RBF)neural network and linear discriminant analysis(LDA)are applied to set up the identification models of places of origin of pepper.The accuracy of SVM model and RBF neural network model is great.The data after pretreatment with db5 wavelet selects only 7 principal components,with the classification accuracy is 100%.The results show that it is feasible to identify the places of origin of pepper based on near infrared spectroscopy,and the pretreatments could improve the accuracy of identification models of the places of origin of pepper effectively.
引文
[1]李鑫,史金儒,羊梦诗,等.胡椒生物碱提取、几何结构以及红外光谱的研究[J].光谱学与光谱分析,2016,36(7):2082-2088.
[2]黄菲菲,吴桂苹,李恒,等.四种方式制备的黑胡椒风味成分分析[J].食品工业科技,2017,38(9):261-267.
[3]郭墨亭,黄雪松.胡椒幼果与黑、白胡椒香气成分的比较研究[J].食品工业科技,2012,33(15):93-100.
[4]Milton C R.Modeling the potential geographic distribution of black pepper(Piper nigrum)in Asia using GIS tools[J].Journal of Integrative Agriculture,2012,11(4):593-599.
[5]Genisheva Z,Quintelas C,Mesquita,et al.New PLS analysis approach to wine volatile compounds characterization by near infrared spectroscopy(NIR)[J].Food Chemistry,2018,246:172-178.
[6]庄新港,王丽丽,吴雪原,等.基于近红外的移动窗口BP神经网络实现山东绿茶产地溯源[J].红外与毫米波学报,2016,35(2):200-205.
[7]李自刚,岳晓禹,李长滨,等.基于变量选择的堆肥胡敏酸含量近红外光谱分析[J].农业机械学报,2017,48(2):300-304.
[8]He W,Zhou J,Cheng H,et al.Validation of origins of tea samples using partial least squares analysis and Euclidean distance method with near-infrared spectroscopy data[J].Spectrochimica Acta Part A:Molecular&Biomolecular Spectroscopy,2012,86(3):399-404.
[9]王元忠,赵艳丽,张霁,等.近红外光谱信息筛选在玛咖产地鉴别中的应用[J].光谱学与光谱分析,2016,36(2):394-400.
[10]吴习宇,祝诗平,黄华,等.近红外光谱技术鉴别花椒产地[J].光谱学与光谱分析,2018,38(1):68-72.
[11]周密,宋哲,龚蕾,等.傅里叶变换红外光谱鉴别芝麻酱中掺杂花生酱[J].中国调味品,2018,43(10):160-163.
[12]陈进,王坤,李耀明.基于Mallat算法的谷物流量信号小波去噪方法[J].农业工程学报,2017,33(3):190-197.
[13]全东平,洪添胜,岳学军,等.基于流形学习算法柑橘叶片高光谱快速分类的研究[C].“农业电气化与信息化工程与学科创新发展”学术年会,2014.
[14]Soleimanzadegan S.Partial least squares method for simultaneous spectrophotometric determination of uracil and 5-fluorouracil in spiked biological samples[J].Journal of Endocrinology,2013(43):111.
[15]祝诗平,梁晶,屠大伟,等.基于近红外光谱与DPLS的潲水油快速鉴别方法[J].西南大学学报(自然科学版),2012,34(5):1-6.
[16]廖秋红,何绍兰,谢让金,等.基于近红外光谱的纽荷尔脐橙产地识别研究[J].中国农业科学,2015,48(20):4111-4119.
[17]侯媛彬,杜京义,汪梅.神经网络[M].西安:西安电子科技大学出版社,2007:125.
[2]黄菲菲,吴桂苹,李恒,等.四种方式制备的黑胡椒风味成分分析[J].食品工业科技,2017,38(9):261-267.
[3]郭墨亭,黄雪松.胡椒幼果与黑、白胡椒香气成分的比较研究[J].食品工业科技,2012,33(15):93-100.
[4]Milton C R.Modeling the potential geographic distribution of black pepper(Piper nigrum)in Asia using GIS tools[J].Journal of Integrative Agriculture,2012,11(4):593-599.
[5]Genisheva Z,Quintelas C,Mesquita,et al.New PLS analysis approach to wine volatile compounds characterization by near infrared spectroscopy(NIR)[J].Food Chemistry,2018,246:172-178.
[6]庄新港,王丽丽,吴雪原,等.基于近红外的移动窗口BP神经网络实现山东绿茶产地溯源[J].红外与毫米波学报,2016,35(2):200-205.
[7]李自刚,岳晓禹,李长滨,等.基于变量选择的堆肥胡敏酸含量近红外光谱分析[J].农业机械学报,2017,48(2):300-304.
[8]He W,Zhou J,Cheng H,et al.Validation of origins of tea samples using partial least squares analysis and Euclidean distance method with near-infrared spectroscopy data[J].Spectrochimica Acta Part A:Molecular&Biomolecular Spectroscopy,2012,86(3):399-404.
[9]王元忠,赵艳丽,张霁,等.近红外光谱信息筛选在玛咖产地鉴别中的应用[J].光谱学与光谱分析,2016,36(2):394-400.
[10]吴习宇,祝诗平,黄华,等.近红外光谱技术鉴别花椒产地[J].光谱学与光谱分析,2018,38(1):68-72.
[11]周密,宋哲,龚蕾,等.傅里叶变换红外光谱鉴别芝麻酱中掺杂花生酱[J].中国调味品,2018,43(10):160-163.
[12]陈进,王坤,李耀明.基于Mallat算法的谷物流量信号小波去噪方法[J].农业工程学报,2017,33(3):190-197.
[13]全东平,洪添胜,岳学军,等.基于流形学习算法柑橘叶片高光谱快速分类的研究[C].“农业电气化与信息化工程与学科创新发展”学术年会,2014.
[14]Soleimanzadegan S.Partial least squares method for simultaneous spectrophotometric determination of uracil and 5-fluorouracil in spiked biological samples[J].Journal of Endocrinology,2013(43):111.
[15]祝诗平,梁晶,屠大伟,等.基于近红外光谱与DPLS的潲水油快速鉴别方法[J].西南大学学报(自然科学版),2012,34(5):1-6.
[16]廖秋红,何绍兰,谢让金,等.基于近红外光谱的纽荷尔脐橙产地识别研究[J].中国农业科学,2015,48(20):4111-4119.
[17]侯媛彬,杜京义,汪梅.神经网络[M].西安:西安电子科技大学出版社,2007:125.