摘要
利用近红外光谱技术快速检测油莎豆乳中蛋白质含量、总固形物含量,实验分别采用凯氏定氮法和考马斯亮蓝法测定蛋白质含量,采用常压干燥法测定总固形物含量;分析比较S-G平滑法、标准正态变量变换(SNV)、多元散射校正(MSC)、一阶微分(FD)、二阶微分(SD)5种光谱预处理方法,利用分段建模法筛选出最佳光谱建模区域,采用留一法交叉验证主成分分析方法确定的主成分数量,结果表明平滑S-G+二阶导数的光谱预处理效果最佳,以1200~1800 nm范围为PLS建模光谱波段,采用考马斯亮蓝法、凯氏定氮法构建的蛋白质含量近红外模型的预测相关系数R分别为0.9755,0.9692,预测均方根误差分别为0.276,0.287;总固形物含量模型的预测相关系数R为0.9699,预测均方根误差为0.326,独立测试集样品的预测值和实测值之间的残差值均较小,接近0,残差和分别为-0.02,0.01,-0.02,研究结果表明建立的近红外光谱快速检测油莎豆乳蛋白质含量、总固形物含量的方法能够满足检测要求,可为油莎豆乳行业健康发展提供技术支撑。
The content of protein and total solids in Cyperus esculentus soymilk is rapidly detected using near-infrared spectroscopy technology.In this experiment,the content of protein is determined by Kjeldahl method and Coomassie brilliant blue staining method respectively,and the content of total solids is determined by atmospheric pressure drying method.Five spectral pretreatment methods including S-G smoothing method,standard normal variable transformation(SNV),multivariate scattering correction(MSC),first order differentiation(FD)and second order differentiation(SD)are analyzed and compared.The segmented modeling method is used to select the optimal spectral modeling region,and the number of principal components is determined by leave-one-out cross validation principal component analysis method.The results show that the spectral pretreatment effect of smoothing S-G+second order derivative is the best,with PLS modeling spectral band in the range of 1200~1800 nm,the predicted correlation coefficient R of the protein content near-infrared model constructed by Coomassie brilliant blue staining method and Kjeldahl method is 0.9755,0.9692 respectively,the predicted root mean square errors are 0.276,0.287 respectively;the predicted correlation coefficient R of the total solid content model is 0.9699,the predicted root mean square error is 0.326.The residual values between the predicted values and the measured values of the independent test set samples are small and close to zero,with the sum of residual values of-0.02,0.01,-0.02 respectively.The research results show that the established rapid detection method for the content of protein and total solids in Cyperus esculentus soymilk by near-infrared spectroscopy can meet the detection requirements,which can provide technical support for the healthy development of Cyperus esculentus soymilk industry.
作者
白瑕
叶春苗
王淼
张卉
李东华
BAI Xia;YE Chun-miao;WANG Miao;ZHANG Hui;LI Dong-hua(School of Pharmaceutical and Bioengineering,Shenyang University of Chemical Technology,Shenyang 110142,China;Liaoyang Vocational College of Technology,Liaoyang 111000,China)
出处
《中国调味品》
CAS
北大核心
2024年第11期158-163,共6页
China Condiment
基金
辽宁省教育厅科学研究经费项目(LJKZ0444,LJKMZ20220785)。
关键词
油莎豆乳
近红外光谱
蛋白质
总固形物
Cyperus esculentus soymilk
near-infrared spectroscopy
protein
total solids
