P43 - Viscosity Analysis of Liquid Food Using Hyperspectral Imaging
- Event
- SMSI 2025
2025-05-06 - 2025-05-08
Nürnberg - Band
- Poster
- Chapter
- Poster Session
- Author(s)
- A. Ghezal, A. König - RPTU University Kaiserslautern Landau, Kaiserslautern (Germany)
- Pages
- 317 - 318
- DOI
- 10.5162/SMSI2025/P43
- ISBN
- 978-3-910600-06-5
- Price
- free
Abstract
This study suggests the use of the third overtone region (675–965 nm) with an NIR hyperspectral camera as a first step toward a non-contact, mobile approach for estimating viscosity in liquid foods—a critical quality parameter. Key spectral features (such as normalized peak intensity, area ratios, FWHM, and peak position) were extracted from selected channel regions and input into a K-Nearest Neighbor (KNR) regression model (k=3), achieving high predictive accuracy (MAE = 0.141 Pa.s, MSE = 0.927 Pa.s², R² = 0.95) and accurately predicting sample viscosity across variations. These findings lay the groundwork for a compact, multi-spectral AMS-based device, supporting in-field quality control and food safety.