近红外特征波长筛选在勾兑梨汁中原汁含量的快速检测中的应用

发布时间：2018-04-26 14:48

  本文选题：近红外 + 特征波长　； 参考：《光谱学与光谱分析》2017年10期

【摘要】：为实现近红外光谱进行勾兑梨汁中原汁含量的快速检测,采用相同可溶性固形物含量的新鲜梨汁和果汁粉冲剂按照原汁含量为0%~100%进行勾兑,并结合遗传算法(GA)、粒子群算法(PSO)以及萤火虫算法(GSOFA)进行特征波长筛选,比较分析四种算法分别建立的偏最小二乘(PLS)模型。结果表明,GA-PLS,PSO-PLS,GSO-PLS,FA-PLS四种模型均能够剔除大部分波长变量,其中以FA-PLS模型效果最佳,不仅保证模型的稳健性,而且简化了模型,提高了预测的精度。为了进一步优选特征波长,利用连续投影算法(SPA)在FA基础上做进一步波长筛选,并比较全波段PLS,SPA-PLS,FA-PLS,FA-SPA-PLS模型,四种模型泛化能力为:FA-PLSPLSFA-SPA-PLSSPA-PLS,其预测均方根误差分别为0.029 1,0.033 3,0.033 9和0.137 0,相应的波长变量数量依次367,765,20和18。其中SPA-PLS波长变量最少,但预测误差远远高于其他三种模型,综合考虑预测精度与波长变量数目,FA-SPA-PLS模型不仅波长变量较少而且预测精度较高,能够有效鉴别勾兑梨汁中原汁含量。研究利用近红外光谱技术为快速鉴别勾兑果汁提供一种有益思路,并通过波长变量筛选简化定量分析模型。
[Abstract]:In order to realize the rapid detection of the content of the juice in the pear juice by near infrared spectrum, the fresh pear juice and the juice powder powder with the same soluble solid content were used in accordance with the content of 0%~100%, and the characteristic wavelength was selected by combining the genetic algorithm (GA), particle swarm optimization (PSO) and the firefly algorithm (GSOFA), and the comparison and analysis of four was compared. The partial least squares (PLS) model is established respectively. The results show that the four models of GA-PLS, PSO-PLS, GSO-PLS and FA-PLS can eliminate most of the wavelength variables, and the FA-PLS model is the best, not only to guarantee the robustness of the model, but also to simplify the model and improve the accuracy of the prediction. The continuation projection algorithm (SPA) performs further wavelength screening on the basis of FA, and compares the full band PLS, SPA-PLS, FA-PLS, and FA-SPA-PLS models. The generalization ability of the four models is: FA-PLSPLSFA-SPA-PLSSPA-PLS, its predicted root mean square error is 0.029 1,0.033 3,0.033 9 and 0.1370 respectively, the number of corresponding wavelength variables in turn 367765,20 and 18. SPA-PLS among them The wavelength variable is the least, but the prediction error is far higher than the other three models. Considering the prediction accuracy and the number of wavelength variables, the FA-SPA-PLS model not only has less wavelength variable and higher prediction precision, and can effectively identify the content of the juice in the pear juice. The quantitative analysis model is simplified by wavelength variable screening.

【作者单位】： 福州大学电气工程与自动化学院;福建省医疗器械和医药技术重点实验室;福州大学生物科学与工程学院;
【基金】：国家自然科学基金项目(61403319) 福建省科技厅国际合作项目(2015I003) 福建省教育厅科技项目(JK2014001)资助
【分类号】：O657.33;TS255.44

【相似文献】

中国期刊全文数据库 前4条

1 樊书祥;黄文倩;李江波;赵春江;张保华;;LS-SVM的梨可溶性固形物近红外光谱检测的特征波长筛选[J];光谱学与光谱分析;2014年08期

2 汪希伟;赵茂程;居荣华;;电可调滤光器光谱图像特征波长的提取方法[J];南京林业大学学报(自然科学版);2014年02期

3 张毅民;白家瑞;刘红莎;汤桂兰;胡彪;;基于近红外特征波长反射率比值法鉴别塑料[J];现代化工;2014年01期

4 刘国海;夏荣盛;江辉;梅从立;黄永红;;一种基于SCARS策略的近红外特征波长选择方法及其应用[J];光谱学与光谱分析;2014年08期

中国重要会议论文全文数据库 前1条

1 屠振华;冯霖;孙丽娟;康颖;籍保平;庆兆s，

本文编号：1806486