Related Experiment Video
Updated: Jun 19, 2026
11:02
Fruit Volatile Analysis Using an Electronic Nose
Published on: March 30, 2012
[Study on Vis/NIR spectra detecting system for watermelons and quality predicting in motion]
Hai-Qing Tian1, Yi-Bin Ying, Hui-Rong Xu
1College of Biosystems Engineering and Food Science, Zhejiang University, Hangzhou 310029, China. hqtian@126.com
Guang Pu Xue Yu Guang Pu Fen Xi = Guang Pu
|October 9, 2009
Summary
This study optimized watermelon quality prediction using Vis/NIR spectroscopy. The Norris differential filtering method effectively reduced motion-induced noise, improving soluble solids content prediction accuracy.
Area of Science:
- Agricultural Engineering
- Spectroscopy
- Food Science
Context:
- Watermelon quality assessment is crucial for the food industry.
- On-line quality prediction requires robust analytical techniques.
- Visible and Near-Infrared (Vis/NIR) diffuse transmittance offers a non-destructive method.
Purpose:
- To adapt Vis/NIR diffuse transmittance for real-time watermelon quality prediction.
- To investigate and mitigate spectral noise caused by dynamic motion.
- To develop and compare statistical models for predicting soluble solids content.
Summary:
- A dynamic spectra detecting system was rebuilt for Vis/NIR diffuse transmittance analysis.
- Least-squares and Norris differential filtering methods were applied to smooth spectra and reduce noise.
- Partial Least Squares (PLS) models were developed to correlate spectral data with soluble solids content.
Impact:
- The Norris differential method significantly improved spectral smoothing and prediction accuracy.
- Achieved high correlation coefficients (r=0.895) and low prediction errors (RMSEP=0.760).
- Demonstrates the potential for accurate, non-destructive quality control of watermelons in motion.
