|
|||||
|
|
| 基于BP神经网络近红外光谱鉴别茶饮料的研究 | |
| 引用本文： | 艾施荣,吴瑞梅,吴燕.基于BP神经网络近红外光谱鉴别茶饮料的研究[J].安徽农业科学,2010,38(14):7658-7659,7662. |
| 作者姓名： | 艾施荣 吴瑞梅 吴燕 |
| 作者单位： | 1. 江西农业大学软件学院,江西南昌,330045 2. 江西农业大学工学院,江西南昌,330045 3. 江西农业大学计算机学院,江西南昌,330045 |
| 基金项目： | 江西省科技厅支撑计划 |
| 摘 要： | 提出了一种快速、准确鉴别茶饮料的新思路。采用美国ASD公司的可见-近红外光谱仪对3种茶原料（龙井茶、乌龙茶和铁观音茶）的饮料进行光谱分析。采用多元散射校正（MSC）方法对样本数据进行预处理,再用主成分分析法提取光谱数据的特征值。通过交互验证确定最佳主成分数为5,作为BP神经网络的输入变量,不同原料茶饮料作为输出变量,建立3层人工神经网络鉴别模型,并用模型对20个预测样本进行预测。模型的回判鉴别率达到100%,模型的预测鉴别率达到98.33%。结果表明,基于BP神经网络的近红外光谱鉴别不同原料茶饮料的方法是可行的。 |
| 关 键 词： | 近红外光谱 BP神经网络 鉴别 茶饮料 |
Discrimination of Different Kinds of Tea Beverage by NIR Spectroscopy Combined to Back Propagation Neural Networks |
|
| AI Shi-rong et al.Discrimination of Different Kinds of Tea Beverage by NIR Spectroscopy Combined to Back Propagation Neural Networks[J].Journal of Anhui Agricultural Sciences,2010,38(14):7658-7659,7662. | |
| Authors: | AI Shi-rong |
| Institution: | AI Shi-rong et al (College of Software Technology,JAU,Nanchang,Jiangxi 330045) |
| Abstract: | A new discrimination method was proposed. Three kinds of tea beverages (mading of Longjin tea,Wulong tea and Tieguanyin tea) were analyzed using near infrared spectroscopy by ASD Company. The sample data were handled through preprocessing method of MSC (multiplicative scatter connection),then eigenvalues were extracted by principal component analysis. The optimal 5 PCs by cross-validation were as in-put in BP neural networks and the best three layer neural networks model was built. The discriminate rate of ... |
| Keywords: | NIR spectroscopy BP neural networks Discrimination Tea beverage |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ | |