Journal: Sensors (Basel, Switzerland)
Article Title: Near-Infrared Hyperspectral Imaging Pipelines for Pasture Seed Quality Evaluation: An Overview
doi: 10.3390/s22051981
Figure Lengend Snippet: Selected applications of NIR-hyperspectral imaging in agriculturally important seeds (reflectance mode).
Article Snippet: Maize seed varietal classification (n = 1632) , PCA; LS-SVM , 400–1000 nm , Wavelength selections: multi-linear discriminant analysis (MLDA) vs. UVE and SPA , ENVI 4.3 , MATLAB 2009b (LS-SVM toolbox) , Full wavelength: calibration of 100% and prediction of 93.26%. Wavelengths, 5–15 selected: MLDA calibration of 99.39–99.88% and prediction of 90.40–93.81%; UVE calibration of 99.46–99.85% and prediction of 88.20–91.94%; SPA: calibration of 98.66–99.40% and prediction of 80.8–87.40% , [ ] .
Techniques: Imaging, Selection, Software, Plasmid Preparation