Nondestructive Testing of Jujube Water Based on the NTRS

【Objective】 Which lays a foundation for the establishment of the water content model of jujube in the next step. 【Methods】 In this experiment, the fruits of Manman jujube and Baodeyou jujube were used as test materials. The water content of jujube was determined by traditional drying method and the moisture content was detected by NIR. 【Results】 The spectra of two varieties were divided into sample sets, and the water content quantitative analysis model was established by using savitzky-Golay smoothing method and partial least square regression analysis method (PLS). A total of 212 samples were obtained, of which 100 and 112 were Manman jujube and Baode jujube respectively, 75 and 84 were random correction models of the two varieties, 25 and 28 were verification models, respectively. The experimental group was divided into two parts: red jujube water content correction model and validation model. The values of SEC (standard deviation of calibration set) and SEP (standard deviation of prediction set) were 1.01% and 1.29% respectively, and the correlation coefficients between calibration set and validation set were 0.878 and 0.883, respectively. 【Conclusion】 S-G smoothing method and PLS regression analysis method are used to establish a water content quantitative detection and analysis model for jujube. The accuracy of water content detection is high, and the cross-correlation coefficient between the real value and the predicted value of water content is higher than 0.850. The difference between the two models is about 0.5%. By analyzing the characteristics of the spectrum, the relationship between the near-infrared spectrum and some quality of jujube is preliminarily established.