Development of a General Prediction Model of Moisture Content in Maize Seeds Based on LW-NIR Hyperspectral Imaging

Moisture content (MC) is one of the important indexes to evaluate maize seed quality. Its accurate prediction very challenging. In this study, long-wave near-infrared hyperspectral imaging (LW-NIR-HSI) system was used, and embryo side (S1) endosperm (S2) spectra each were extracted, as well average spectrum (S3) both being calculated. The partial least square regression (PLSR) least-squares support vector machine (LS-SVM) models established. uninformative variable elimination (UVE) successive projections algorithm (SPA) employed reduce complexity models. results indicated that S3-UVE-SPA-PLSR S3-UVE-SPA-LS-SVM achieved best accuracy with an RMSEP 1.22% 1.20%, respectively. Furthermore, combination (S1+S2) S1 S2 also used establish obtain a general model. S1+S2-UVE-SPA-LS-SVM model more valuable Rpre 0.91 1.32% for MC prediction. This can decrease influence different input (i.e., or S2) on performance. overall study LW-HSI technology combined could realize non-destructive stable in seeds.