Detection of soluble solid content in table grapes during storage based on visible-near-infrared spectroscopy

The soluble solid content (SSC) in grapes significantly influences their flavour and plays an integral role in evaluation of the quality and consumer acceptance. This study employed visible near-infrared (Vis-NIR) spectroscopy to rapidly quantify SSC in table grapes during storage. A predictive model was developed to construct a correlation between the spectral data and the measured SSC, while a comparative analysis was undertaken to assess the effects of various spectral preprocessing techniques. Successive projection algorithms (SPA), uninformative variable elimination (UVE), and the competitive adaptive reweighting algorithm (CARS) were adopted to eliminate redundant variables from both the original and preprocessed spectral data. The partial least squares regression (PLSR), and support vector regression (SVR) algorithms were adopted to establish a predictive model. Comparing the modelling results derived from whole-band spectral data with those obtained from selected spectral variables, the optimal spectral prediction model was formulated utilizing PLSR. The model, which incorporated filtered characteristic wavelength spectral data obtained through CARS following standard normal variate (SNV) preprocessing yielded optimum results with the correlation coefficients of the calibration set (RC), and the prediction set (RP) were 0.956 and 0.940, respectively. The root mean square errors of the calibration set (RMSEC), and prediction set (RMSEP) were 0.683 and 0.769, respectively, while the ratio of prediction to deviation (RPD) was 2.899. These results suggest that the application of Vis-NIR spectroscopy technology could effectively detect the SSC in grapes during storage, and it can provide a valuable reference for the rapid assessment of the table grape quality.