Application of hyperspectral imaging to discriminate waxy corn seed vigour after aging

Abstract

A hyperspectral imaging system covering 400 - 1000 nm spectral range was applied for vigour detection of waxy maize seeds after artificial aging. After spectral pre-processing, the characteristic wavelength was selected by uninformative variable elimination (UVE), competitive adaptive reweighted sampling (CARS), and random frog (RF) methods. The moisture, starch, protein, and fat contents were measured for each grade of seed, and these values were correlated with the spectrum. Finally, the vitality detection model was established by least squares support vector machine (LS-SVM), extreme learning machine (ELM), and random forest (RF). The prediction sets exhibited high classification accuracy (> 99%) for 115 features. The model constructed from the bands significantly correlated with chemical composition (CC), and was better than the classic feature selection methods. The overall results indicated that hyperspectral imaging could be a potential technique to assess seed vigour.