A virtual seed file: the use of multispectral image analysis in the management of genebank seed accessions

Michael Adsetts Edberg Hansen1, Fiona R. Hay2* and

Jens Michael Carstensen1,3

1Videometer A/S, DK-2970 Hørsholm, Denmark, 2International Rice Research Institute, Los Ban˜ os, Laguna, Philippines and 3Technical University of Denmark, DK-2800 Lyngby, Denmark

Plant Genetic Resources: Characterization and Utilization; 1–4

Abstract

文摘

We present  a method  for multispectral seed phenotyping as a fast and robust tool for managing genebank accessions.  A multispectral vision system was used to take images of the seeds of 20 diverse varieties of rice (approximately 30 seeds for each variety). This was followed by extrac- tion of feature information  from the images. Multivariate analysis of the feature data was used to classify seed phenotypes according  to accession.  The proportion of correctly classified rice seeds was 93%. We conclude  that the multispectral image analysis could play a role in compar- ing incoming seeds against existing accessions, identifying different seed types within a sample of seeds  and/or in checking  whether  regenerated seeds  match the original seeds.

    我们提供了一种快速、功能强大的多光谱种子表型成像的工具用以进行基因库管理。利用多光谱成像系统拍摄了20个不同品系种子图像（每个品系约需要30粒种子）。之后从图像中提取特征信息。多品系分析的特征数据用来对表型分类。水稻种子分类正确率达到93%。因而我们推论，种子成像系统分析可在用来对新添加种子与现有种子进行比较，在种子样品中识别不同种子类型，并用来检测再繁育种子是否与先前保存种子一致。

Keywords:      genebank management; multispectral  image analysis; rice; seed  phenotype