基于可见近红外的柿饼涩味定量定性分析  

Quantitative and qualitative analysis of astringency in dried persimmon based on near infrared spectroscopy

作　　者：陈超锋 黎庆涛[1] 韦保耀[1] 夏宁[1] 王海军 滕建文[1] CHEN Chaofeng;LI Qingtao;WEI Baoyao;XIA Ning;WANG Haijun;TENG Jianwen(Guangxi University,Nanning 530004;Institute of Quality Standard and TestingTechnology for Agro-Products,Guangxi Academy of Agricultural Sciences,Nanning530004)

机构地区：[1]广西大学,南宁530004 [2]广西壮族自治区农业科学院农产加工质量安全与检测技术研究所,南宁530004

出　　处：《食品科技》2019年第8期322-328,共7页Food Science and Technology

基　　金：广西创新驱动发展专项(桂科AA17204038);广西农业科学院科技发展基金项目(桂农科2018JZ11)

摘　　要：柿饼涩味评价是柿饼生产流通环节的重要控制工作。目前没有建立柿饼的快速无损检测方法。文章以恭城月柿为原料,建立涩味可见近红外快速无损检测模型。结果表明:柿饼水分含量在32.1%~36.17%可见近红外定量分析中,发现在460~1050nm与1300~1680nm波段范围内,采用改进偏最小二乘回归算法、二阶导数结合标准正常化处理(Standard normal variate,SNV)的建模效果最好。其定标交互验证相关系数(Correlation coefficient of cross validation,1-VR)和预测相关系数(Correlation coefficient of prediction,R2p)分别为0.878和0.865,定标交互验证均方根误差(Root mean standard Error of cross validation,RMSECV)和预测均方根误差(Root mean square error of prediction,RMSEP)分别为0.105、0.125g/100g.d;定性模型中,柿饼水分含量在33.2%~36.11%范围内采用450~1050nm与1300~1650nm波段结合去散射处理(Detrendonly,D)、一阶导数预处理方法最好。判别模型正确率93.1%,预测正确率为72.22%~88.89%。因此,近红外光谱技术可用于柿饼涩味快速无损的定量定性分析。Astringency assessment was an important control of the production and distribution of dried persimmons.There was currently no rapid non-destructive testing method of dried persimmons.In this paper,Yue persimmon was used as raw material to develop a predictive model for soluble tannin content and astringent classification near-infrared.The moisture content of near-infrared quantitative modeling was between 32.1% and 36.17%.The results showed that in the range of 460~1050 nm and 1300~1680 nm, the modified partial least squares (MPLS) model,with the second derivative and standard normalization (SNV) only,provided better prediction performance for the soluble tannin content of astringent persimmon fruit,with the correlation coefficient of cross validation of calibration (1-VR) and correlation coefficient of prediction (R2 p),the root mean square error of cross validation of calibration (RMSECV) and the root mean square error of prediction (RMSEP) of 0.878,0.865,0.105,0.125 g/100 g.d.Qualitative model,the persimmon moisture content in the range of 33.2%~36.11% using 450~1050 nm and 1300~1650 nm band combined with detrend treatment,first derivative pretreatment method was best.The correct rate of the discriminant model was 93.33%,and the prediction accuracy rate was 72.22%~88.89%.Therefore,nearinfrared spectroscopy can be used for quantitative and qualitative analysis of the astringency assessment of dried persimmons.

关 键 词：柿饼 可见近红外光谱 可溶性单宁 定量定性 

分 类 号：O657.33[理学—分析化学]