Quantitative determination of rice starch based on hyperspectral imaging technology

Abstract

In this study, a method for the quantitative determination of rice starch based on hyperspectral imaging technology was proposed. First, the hyperspectral imaging system in the spectral range of 871–1766 nm was used to collect the hyperspectral images of 100 rice samples of 10 starch grades. The support vector regression (SVR) model was established to determine the starch content by using full-wavelength spectra data. Among all the models, the SVR-principal component analysis (SVR-PCA) model with the Radial Basis Function showed the best results. To simplify the calibration model, PCA was used for feature extraction and the cumulative contribution rate of the first six principal components reached 99%, which could reflect most of the information of the full spectra data. Three new regression models based on the selected wavelengths were developed and the results were improved obviously. The SVR-PCA model obtained the best accuracy in prediction and calibration with the determination coefficients of prediction (R2p) of 0.991, root mean square error of prediction (RMSEP) of 0.669%, the determination coefficients of calibration (R2c) of 0.989, and root mean square error of calibration (RMSEC) of 0.445%. The overall results from this study demonstrated that the hyperspectral image technology is feasible to detect rice starch.