Cultivars identification of oat ( Avena sativa L.) seed via multispectral imaging analysis.

Abstract

Cultivar identification plays an important role in ensuring the quality of oat production and the interests of producers. However, the traditional methods for discrimination of oat cultivars are generally destructive, time-consuming and complex. In this study, the feasibility of a rapid and nondestructive determination of cultivars of oat seeds was examined by using multispectral imaging combined with multivariate analysis. The principal component analysis (PCA), linear discrimination analysis (LDA) and support vector machines (SVM) were applied to classify seeds of 16 oat cultivars according to their morphological features, spectral traits or a combination thereof. The results demonstrate that clear differences among cultivars of oat seeds could be easily visualized using the multispectral imaging technique and an excellent discrimination could be achieved by combining data of the morphological and spectral features. The average classification accuracy of the testing sets was 89.69% for LDA, and 92.71% for SVM model. Therefore, the potential of a new method for rapid and nondestructive identification of oat cultivars was provided by multispectral imaging combined with multivariate analysis.

Plant phenotyping relevance

オート種子の形態・スペクトル特徴をマルチスペクトル画像と多変量解析で取得・分類する手法が研究の中心であり、植物生殖資材の表現型・品種識別に直接関係するため。

Code and data availability

The article reports oat seed multispectral imaging and LDA/SVM classification, but no public phenotype dataset, images, or author analysis code is deposited. The data availability statement only offers inquiries to the corresponding author, and the supplementary material is not described as containing raw data or code.

No evidence-backed public reproduction asset is currently recorded.