Seed Vigour of Stored Seeds

Purpose

We aim to find a correlation between the spectral properties of a fresh seed and the seedling quality that originates from it after a period of storage, in order to predict storability of individual seeds. This information is used to design a method to sort seeds into classes with uniform seed vigour and storability. The ultimate aim is to predict seed storage potential to optimise utilisation of this aspect. This leads to an efficient use of produced seed batches with minimum losses.

Approach

The approach of this project is to individually measure the seed quality in a non-destructive way, using the different sensors of the iXeed DataCollector such as X-ray, chlorophyll fluorescence, NIR, etc. Cabbage seeds are used as a model. After analysis, the seeds are individually stored for a longer period of time and we end up with aged seeds of which the properties in fresh condition are known. These aged seeds are germinated and the resulting seedlings are individually measured using the Marvin 3D phenotyping machine. Next, artificial intelligence is used to find the most predictive properties for predicting seed vigour.

Of course, aged seeds are required at an early stage in the project to develop the AI methods. For this, we artificially age fresh seeds using experimental ageing under elevated oxygen pressure, the EPPO technique which was developed at WUR.

Initial results

So far, the AI approach pursued, appears to give good results and an initial estimation of needed seeds for a reliable prediction has been made. The natural ageing experiment has been started for evaluation at the end of the project. Potentially interesting experiments concerning the volatile organic compounds emitted by seeds, and Raman spectroscopy have started for additional information on seed quality during ageing.