Prediction of Soluble Solids Content and Firmness of Pears Using Hyperspectral Reflectance Imaging

Abstract

Hyperspectral imaging technique was investigated to determine the soluble solids content (SSC) and firmness of pears. A total of 160 pear samples were prepared for the calibration (n = 120) and prediction (n = 40) sets. A hyperspectral imaging system was used to acquire hyperspectral reflectance image from each pear in visible and near infrared (400–1000 nm) regions. Mean spectra were extracted from the regions of interest for the hyperspectral image of each pear. Spectral data were first pretreated with different preprocessing techniques and analyzed using partial least square (PLS) to establish calibration models. However, the large size of spectral data contains a large number of redundant variables that lead to complexity and poor predicting ability of calibration models. Several variable selection methods were investigated to select effective wavelength variables for the determination of SSC and firmness of pear. In this study, the variables selected by successive projection algorithm (SPA), competitive adaptive reweighted sampling (CARS) and the combination of CARS and SPA were used for PLS regression. The CARS-SPA-PLS models based on 25 and 22 variables achieved the optimal performance for two internal quality indices compared with full-spectrum PLS, CARS-PLS, and SPA-PLS models. The correlation coefficient (r pre) and root mean square error of prediction (RMSEP) by CARS-SPA-PLS were 0.876, 0.491 for SSC and 0.867, 0.721 for firmness, respectively. The overall results indicated that the CARS-SPA was a powerful way for the selection of effective variables and the hyperspectral imaging system together with CARS-SPA-PLS model could be applied as a fast and potential method for the determination of SSC and firmness of pear.