Towards numerical practices in variety testing: a rationale to select the most promising traits

Hadhami Garbouge, Nicolas Mascher, Joseph Peller, Philippe Vermeulen, Didier Demilly, Rick Van de Zedde, Cécile Collonnier, Mark Mueller-Linow and David Rousseau

UPOV, Technical Working Party on Automation and Computer Programs

Abstract
In this opinion paper, we focus on variety testing offices in the European Union. Variety testing nowadays is mainly done with traditional visual inspections. As a consequence, introducing automation and sensors in variety testing could lead to an increase of throughput together with a higher reproducibility. However the task is huge since there are thousands of characteristics to be measured within variety testing. We propose a rationale to select the characteristics which may be the most promising to benefit from an automation via numerical practices. An important criterion for variety testing is that the measurement have to be achieved with low-cost sensors. For illustration, we describe a strategy to select the most promising characteristics that we believe could benefit from such an integration with low-cost sensors for a small set of crops of major interest in food production.

Link: https://www.upov.int/meetings/en/doc_details.jsp?meeting_id=55673&doc_id=511975