Machine vision based alternative testing approach for physical purity, viability and vigour testing of soybean seeds (Glycine max).

The conventional methods for seed quality testing have several limitations as they involve visual assessment and are destructive. In this context, a study was performed to assess the suitability of non-contact, non-destructive type imaging techniques such as visible imaging and X-ray imaging for conducting physical purity, viability and vigour tests of soybean seeds. The seeds that appeared healthy in external surface examination using visible tests as well as in internal assessment using X-ray tests were classified as sound seeds while the other seeds were marked as not-sound seeds. The obtained results were then correlated with the results of the standard germination tests. The high correlation results between the imaging tests and the standard conventional germination tests indicate the effectiveness and usability of the proposed image analysis based technique as an attractive alternative to the existing quality assessment methods for soybean seeds.