AUTHOR=Wang Zheli , Huang Wenqian , Tian Xi , Long Yuan , Li Lianjie , Fan Shuxiang 

TITLE=Rapid and Non-destructive Classification of New and Aged Maize Seeds Using Hyperspectral Image and Chemometric Methods

JOURNAL=Frontiers in Plant Science

VOLUME=Volume 13 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2022.849495

DOI=10.3389/fpls.2022.849495

ISSN=1664-462X

ABSTRACT=The aged seeds have a great influence on seed vigor and corn growth. Therefore, it is very important for the planting industry to identify aged seeds. In this study, hyperspectral reflectance imaging (1000-2000 nm) was employed for identifying aged maize seed using harvested seeds in different years. The average spectra of embryo side, endosperm side and both sides were extracted. Support vector machine (SVM) algorithm was used to develop classification models based on full spectra to evaluate the potential of hyperspectral imaging for maize seed detection. Using the Principal component analysis (PCA) and ANOVA to reduce data dimensionality and extract feature wavelengths. The classification models obtained the prefect performance based on full spectra with the accuracy of 100%. The performance of models established with the first three principal components were similar to full spectra models, but that of PCA loading models became worse. Compared with other spectra, the two-band ratio (1987 nm/1079 nm) selected by ANOVA form embryo side spectra obtained better classification accuracy of 95.00% for prediction set. The image texture features including histogram statistics (HS) and gray level co-occurrence matrix (GLCM) were extracted from the two-band ratio image to establish fusion models. The results demonstrated that two-band ratio selected from embryo side spectra combined with image texture features achieved the classification of maize seeds harvested in different years with the accuracy of 97.5% for prediction set. The overall results indicated that two wavelengths combined with image texture features could detect aged maize seeds effectively. The proposed method was conducive to development of multi-spectral detection equipment.