花生籽仁食用感官品质近红外分析模型构建

Construction of Near Infrared Spectroscopy Models on Prediction of Eating Quality of Peanut Kernel

采集31份日光干燥的花生籽仁的近红外光谱,并进行生花生仁和烤花生仁感官品质评价。采用交叉检验,首次构建花生籽仁感官品质近红外模型。经优化,生花生仁脆性、甜度和细腻度,烤花生仁脆性和细腻度,以上最佳光谱预处理方法均为“一阶导数+多元散射矫正”。生花生仁脆性近红外模型R~2为83.6,RMSECV为0.227;生花生仁甜味近红外模型R~2为89.02,RMSECV为0.137;生花生仁细腻度近红外模型R~2为90.92,RMSECV为0.102。烤花生仁脆性近红外模型R~2为88.29,RMSECV为0.161;烤花生仁细腻度近红外模型R~2为76.97,RMSECV为0.154。花生感官品质近红外模型的应用,将会对食用花生感官品质育种发挥积极的促进作用。

Near infrared spectra of 31 sundried peanut kernel samples were collected and evaluated for sensory quality of raw and roasted peanuts.Near infrared spectroscopy(NIR)models for sensory quality of peanut kernels were developed using cross-validation strategy for the first time.After optimization,the best spectral pretreatment method was first-order derivative+multiple scattering correction for the crispness,sweetness and fineness of raw peanut kernels,and the crispness and fineness of roasted peanut kernels.To raw peanut kernels,the NIR model for crispness had a R~2 of 83.6 and a RMSECV of 0.227;the NIR model for sweetness had a R~2 of 89.02 and a RMSECV of 0.137;and the NIR model for finesse had a R~2 of 90.92 and a RMSECV of 0.102.For roasted peanut kernels,the NIR model for crispness possessed a R~2 of 88.29 and a RMSECV of 0.161,and the NIR model for finesse owned a R~2 of 76.97 and a RMSECV of 0.154.The application of the NIR models for peanut sensory quality will play a positive role in accelerating edible peanut breeding for sensory quality.