The potential application of HSI and VIS/NIR spectroscopy for non-invasive detection of starch gelatinization and head rice yield during parboiling and drying process

Abstract

Real-time detection of starch gelatinization (SG) and head rice yield (HRY) during parboiling is crucial for reducing product loss in the rice industry. Experimental measurement of SG and HRY is costly and time-consuming, making real-time assessment challenging. This study employed hyperspectral imaging (HSI) and VIS/NIR spectroscopy (VNS) to predict these parameters and identify effective wavelengths. Rice samples were processed at varying soaking (60–70 ℃) and drying (50–70 ℃) conditions using a hybrid infrared-convective-microwave (ICM) dryer. Spectral data were preprocessed using five different methods, with the Savitzky–Golay method proving to be the most effective. A combination of a Decision Tree and a Learning Automata metaheuristic algorithm (DT-LA) emerged as the best approach for identifying the most effective wavelengths. Subsequently, regression models based on Partial Least Squares Regression (PLSR) and an Artificial Neural Network (ANN) were developed to predict SG and HRY. The ANN demonstrated superior predictive accuracy, achieving R² values of 0.99 for SG and 0.98 for HRY. Additionally, HSI data yielded slightly better results than VNS data. The findings revealed significant increases in SG levels (from 8.78 % to 88.51 %) and HRY (from 45 % to 70.35 %) associated with higher temperatures and longer soaking times, as well as increased soaking temperatures and drying times when using a hybrid ICM dryer except for samples soaked at 65 °C, where those samples exhibited the highest HRY. These results support the development of smart dryers and optimized parboiling processes to enhance rice yield and improve production efficiency.