摘要
为实现近红外光谱技术对新鲜辣椒果实中辣椒素含量的准确预测,以杭椒类辣椒为研究对象,采集新鲜辣椒果实的近红外光谱信息,结合高效液相色谱法,分别采用连续投影算法(successive projection algorithm,SPA)、竞争性自适应重加权采样法(competitive adaptive reweighted sampling,CARS)、无信息变量消除法(uninformation variable elimination,UVE)提取特征波长,建立偏最小二乘法(partial least squares,PLS)预测模型,并比较了全谱建模与特征波长建模的预测效果和运算效率。结果显示:CARS-PLS模型的预测效果最好,预测集相关系数和均方根误差分别为0.8386和0.0148 mg/g。此外,与全谱建模的输入变量200相比,基于CARS选择的特征波长建模的输入变量减少了96%,这说明运用特征波长建模大大地简化了模型,提高了运算效率。本试验表明,基于特征波长建模的近红外光谱技术对于新鲜辣椒果实中辣椒素含量的检测是可行的。
In order to investigate the potential of near-infrared spectroscopy for accurately predicting the capsaicin content in fresh chili peppers,taking Hangzhou chili pepper as a material,the near-infrared spectroscopy was employed to acquire spectral information of chili peppers,and high-performance liquid chromatography was conducted to obtain the reference values of capsaicin content.Three different variable selection methods with successive projection algorithm(SPA),competitive adaptive reweighted sampling(CARS)and uninformation variable elimination(UVE)were performed to select the optimal wavelengths.Partial least square(PLS)models based on full spectra and optimal wavelengths were then developed to predict the capsaicin content,and the prediction performances and operation efficiency were compared.The results showed that the CARS-PLS model yielded the best prediction performances,with the correlation coefficient of 0.8386 and root-mean-square error of prediction set of 0.0148 mg/g.In addition,compared with the full spectra of 200 wavelengths,the number of the optimal wavelengths selected by CARS was reduced by 96%,which indicated that optimal wavelengths can be used to simplify the models and improve the operation efficiency.The above results demonstrate that the near-infrared spectroscopy based on optimal wavelengths is feasible for the detection of capsaicin content.
作者
吕晓菡
蒋锦琳
杨静
陈建瑛
岑海燕
傅鸿妃
周毅飞
LÜXiaohan;JIANG Jinlin;YANG Jing;CHEN Jianying;CEN Haiyan;FU Hongfei;ZHOU Yifei(Hangzhou Academy of Agricultural Sciences,Hangzhou 310024,China;College of Biosystems Engineering and Food Science,Zhejiang University,Hangzhou 310058,China;College of Agriculture and Food Science,Zhejiang A&F University,Hangzhou 311300,China)
出处
《浙江大学学报（农业与生命科学版）》
CAS
CSCD
北大核心
2019年第6期760-766,共7页
Journal of Zhejiang University:Agriculture and Life Sciences
基金
浙江省农业（蔬菜）新品种选育重大专项（2016C02051）
杭州市农业科研主动设计项目（20162012A01）
关键词
近红外光谱技术
辣椒素
辣椒
特征波长
near-infrared spectroscopy
capsaicin
chili pepper
optimal wavelengths
