Determine Reducing Sugar Content in Potatoes Using Hyperspectral Combined with VISSA Algorithm

摘要

In order to explore nondestructive and rapid detection of reducing sugar in potatoes, hyperspectral imaging technology was applied for quantitatively analyze reducing sugar in potatoes. A quantitative analysis model of reducing sugar in potatoes was constructed by partial least squares method. Sacitzky-Golay (SG) smoothing filter, standard normal variable transformation (SNV), first derivative (FD), multivariate scattering correction (MSC) and other optimization models were used. Variable Iterative Space Shrinkage algorithm (VISSA) is proposed for feature wavelength selection, and compared with competitive adaptive weighting algorithm (CARS). A total of 229 samples were prepared, and the SXYP method was used to divide the samples. 181 samples were selected as the correction set and the remaining 48 samples as the verification set. The results showed that, the model of reducing sugar content in potato spectrum pretreated by SG + SNV was the best, and the partial least squares regression model (VISSA-PLS) based on VISSA algorithm to select characteristic variables had good prediction ability. The determination coefficient of model validation set was 0.8144, and the root mean square error of validation set was 0.0238. It was concluded that the model has good predictive performance after optimization and achieves rapid and nondestructive detection of reducing sugar in potatoes.