FOOD SCIENCE ›› 2019, Vol. 40 ›› Issue (10): 285-291.doi: 10.7506/spkx1002-6630-20180508-119

• Safety Detection • Previous Articles     Next Articles

Nondestructive Detection of Sucrose Content of Lingwu Changzao Jujubes by Hyperspectral Imaging

CHENG Lijuan1, LIU Guishan1, HE Jianguo1,*, YANG Xiaoyu1, WAN Guoling1, ZHANG Chong1, MA Chao2   

  1. 1. Non-Destructive Detection Laboratory of Agricultural Products, School of Agriculture, Ningxia University, Yinchuan 750021, China; 2. School of Physics and Electronic-Electrical Engineering, Ningxia University, Yinchuan 750021, China
  • Online:2019-05-25 Published:2019-05-31

Abstract: In the present study, we focused on developing a rapid non-destructive method for rapid detection of the sugar content in Lingwu Changzao jujubes in order to predict the distribution of sugar content. High performance liquid chromatography (HPLC) was used to detect the sucrose content. Simultaneously, hyperspectral imaging combined with chemometrics method was used to establish a predictive model for sucrose content. The predictive models were built based on the full spectra and the feature wavelengths using partial least squares regression (PLSR), principle component regression (PCR) and multi-variable linear regression (MLR) through dimensionality reduction using competitive adaptive reweighed sampling (CARS), successive projections algorithm (SPA) and uninformative variable elimination (UVE). It was found that the correlation coefficient was increased from 0.611 to 0.846 by removing the abnormal samples using the Monte Carlo method; orthogonal signal correction (OSC) was the optimal preprocessing approach, and the correlation coefficients of calibration and prediction sets (RC and RP) of the PLS model were 0.853 and 0.794, respectively; SPA, UVE, CARS, CARS + SPA and CARS + UVE methods were used to select 5, 21, 17, 10, and 18 characteristic wavelengths, respectively. The CARS-PCR model was the best among the models developed, and its RC, RP, root mean square error of calibration (RMSEC) and root mean square error of prediction (RMSEP) values were 0.861, 0.843, 0.013 mg/g and 0.014 mg/g, respectively.According to the results of this study, hyperspectral imaging can be used to predict the sucrose content of jujube, laying the foundation for insights into the internal quality of jujubes.

Key words: Lingwu Changzao jujube, sucrose, visible-near infrared spectroscopy, high performance liquid chromatography (HPLC)

CLC Number: