Design and Experiment of Non-destructive Testing System for Moisture Content of In-situ Maize Ear Kernels Based on VIS-NIR
Yeyuan Jiang,Dongxing Zhang,Li Yang,Tao Cui,Xiantao He,Duoyang Wu,Jiaqi Dong,Chuan Li,Shulun Xing
DOI: https://doi.org/10.1016/j.jfca.2024.106369
IF: 4.3
2024-01-01
Journal of Food Composition and Analysis
Abstract:To facilitate rapid, non -destructive, cost-effective continuous detection of Moisture Content in corn kernels, a Near-infrared (VIS-NIR) spectroscopy based in -situ maize ear moisture detection device was developed, utilizing machine learning for predictive modeling. Field experiments (30-35 degrees C) assessed three preprocessing algorithms: z -score normalization (ZS), Orthogonal Signal Correction (OSC), and a ZS-OSC combination, with ZS-OSC selected for its superior performance (R-2 >= 0.90, RMSE <= 2.12 %, RPD>2.9). Spectral imaging from 410 to 940 nm was used to develop moisture prediction models via Partial Least Squares Regression (PLSR) and Support Vector Machine (SVM), where PLSR is suited for single variety (R2 >= 0.82, RMSE <= 2.62 %, RPD >= 2.2) and SVM for both single and mixed varieties. Additionally, grain temperature's impact on model performance was analyzed, showing decreased accuracy across temperatures of 30-35 degrees C, 35-40 degrees C, and 40-45 degrees C. The final device and models excelled in 30-35( degrees)C field tests, achieving R2 >= 0.88, RPD>2.5, RMSE <= 0.901 %, with less than 1.82 % deviation between predicted and actual values, and all classification indices over 84.38 %. The device is proven accurate and effective for corn grain moisture detection, offering valuable insights for in -situ maize moisture content analysis.