Nondestructive measuring soluble solid contents and weight of intact pears based on on-line near-infrared spectroscopy

摘要

The objective of this study was to investigate the potential of MRS as an accurate, reliable and nondestructive method to measure SSC and weight parameters of intact pears. The effects of fruit moving speed on SSC and weight prediction were evaluated. Diffuse reflectance spectra of 200 samples were collected by a fiber spectrometer in the range of 840~950nm at 5 different speeds: 2, 3, 4, 5 and 6 fruits per second. One-way analysis of variance (ANOVA) method was employed to assess effect of speed with partial least squares regression (PLSR). At the speed of 4 fruits per second, five types of data pretreatment were adopted before calibration. Multiple linear regression (MLR), principal component regression (PCR) and PLSR models were developed in the calibration set (n=150). 50 unknown samples were applied to evaluate the performance of the models. SSC was measured with a root mean squares error of prediction (RMSEP) comprised between 0.60 and 0.88 °Brix and correlation coefficient (r) between 0.57 and 0.85. As far as weight was concerned, the best results were obtained by PLSR combined with smoothing spectra (r=0.93, RMSEP=14.29 g). The results demonstrated that on-line NIRS was feasible for measuring SSC and weight of intact pears nondestructively.