摘要
采用近红外光谱技术结合偏最小二乘法建立预测红薯淀粉及全粉粉丝中薯粉含量的定量模型,实现薯粉含量的在线快速检测。分别制作180份红薯淀粉粉丝和红薯全粉粉丝样品,以一阶导数、最大最小归一化处理、移动平均平滑、多元散射校正、标准正态化等计量学方法预处理光谱。结果表明,选择波数9403.6~7498.2 cm^(–1)、6101.9~4246.7 cm^(–1)+消除常数偏移量所建的红薯淀粉粉丝模型效果最好,波数9403.6~4597.6 cm^(–1)+减去一条直线所建的红薯全粉粉丝模型效果最好,预处理后2个模型相关系数分别为0.9875和0.9892,交叉验证均方根误差(RMSECV)分别为1.23和1.13,校正后预测相对分析偏差(RPD)分别为6.83和7.42。外部验证预测相关系数为0.9625和0.9714,相对标准偏差(RSD)均小于1,模型具有较高的准确度。近红外光谱技术可以实现贵州红薯粉丝中薯粉含量的快速检测。
The quantitative models for predicting the content of sweet potato flour in sweet potato starch and whole flour vermicelli were established by using near-infrared spectroscopy technology combined with partial least squares method,thereby realizing rapid online detection of sweet potato flour content.Total 180 samples of sweet potato starch vermicelli and whole sweet potato vermicelli were prepared,respectively.The spectra was preprocessed by metrological methods,including first derivative,maximum and minimum normalization,moving average smoothing,multiple scattering correction,standard normalization,etc.The results showed that the best sweet potato starch vermicelli model was established by choosing wave numbers of 9403.6~7498.2 cm^(-1),6101.9~4246.7 cm^(-1) combined with the constant offset eliminated.Meanwhile,the best sweet potato whole flour vermicelli model was selected with the wave numbers of 9403.6~4597.6 cm^(-1) combined with minus a straight line.The correlation coefficients of the two models were 0.9875 and 0.9892,respectively.The root means square error values (RMSECV) were 1.23 and 1.13,respectively.The adjusted relative analysis deviation values (RPD) were 6.83 and 7.42,respectively.Furthermore,the externally verified prediction correlation coefficients were 0.9625 and 0.9714,respectively,and the relative standard deviation (RSD) was less than 1,indicating that the model possessed high accuracy.Near-infrared spectroscopy technology can quickly detect the content of sweet potato flour in Guizhou sweet potato vermicelli.
作者
潘牧
吕都
刘辉
石义权
刘永翔
黄珊
袁再敏
李俊
PAN Mu;LYU Du;LIU Hui;SHI Yiquan;LIU Yongxiang;HUANG Shan;YUAN Zaimin;LI Jun(Biological Technology Institute of Guizhou Academy of Agricultural Sciences,Guiyang 550006,China;Guizhou Shenkang Original Ecological Food Co.,Ltd.,Qiannan 550600,China;Guizhou Xinqidian Ecological Agricultural Science and Technology Development Co.,Ltd.,Guiyang 550025,China;Guizhou Key Laboratory of Biotechnology,Guiyang 550006,China)
出处
《食品科技》
CAS
北大核心
2022年第4期316-321,共6页
Food Science and Technology
基金
贵州省省级科技计划项目(黔科合支撑[2021]一般174,黔科合成果[2019]4243号,黔科合成果[2020]1Y023号)
贵州省甘薯工程技术研究中心项目(黔科合平台人才[2019]5201号)。
关键词
近红外光谱技术
偏最小二乘法
快速检测
红薯粉丝
红薯全粉
near infrared spectroscopy
partial least squares method
rapid detection
sweet potato vermicelli
whole sweet potato flour
