Evaluation of quality attributes and milling metrics of glutinous rice stored under different storage conditions using infrared thermal imaging
Infrared thermal imaging (ITI) has emerged as a promising tool for non-destructive evaluation of agricultural product quality. This study investigates the application of ITI for monitoring the quality attributes and milling metrics of glutinous rice (GR) during storage under varying conditions. GR s...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | en |
| Published: |
Elsevier
2026
|
| Subjects: | |
| Online Access: | http://psasir.upm.edu.my/id/eprint/123066/1/123066.pdf http://psasir.upm.edu.my/id/eprint/123066/ https://www.sciencedirect.com/science/article/pii/S0022474X26000135 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Infrared thermal imaging (ITI) has emerged as a promising tool for non-destructive evaluation of agricultural product quality. This study investigates the application of ITI for monitoring the quality attributes and milling metrics of glutinous rice (GR) during storage under varying conditions. GR samples were first dried at 50 °C, 60 °C, and 70 °C, then stored at freezing (−10 °C), cold room (6 °C), and ambient (26 °C) temperatures for 6 months (24 weeks). Thermal images (TI) of the GR were acquired biweekly to evaluate the changes in key quality indicators, including moisture content (MC), germination growth rate (GGR), water absorption capacity (WAC), whiteness index (WI), head rice yield (HRY), and broken rice yield (BRY). Relevant features were extracted from the TI and correlated with the corresponding physicochemical properties of the GR. Ten machine learning (ML) algorithms were tested to predict the quality attributes of GR based on the features extracted from TI data. Among them, the ET model demonstrated superior performance compared to the others. Its predictive capability was further enhanced through grid search (GS) hyperparameter tuning (HPT), achieving an R2 of 0.939 and an RMSE of 0.178 for MC. The accuracy and reliability of the ET model were further supported by parity plots. Overall, the findings underscore the potential of integrating ITI with ML for non-destructive, real-time monitoring of GR quality, offering valuable insights for optimizing postharvest handling and storage practices. |
|---|