Computer vision: speeding up varietal selection
Improving a plant variety calls for thousands of precise, repetitive measurements. Computer vision automates them and makes R&D data more reliable.
Illustration generated by artificial intelligence (OpenAI), 10 August 2026 — chicory roots and a food-fibre sample. It shows no real person, brand or signage, and depicts no site or facility belonging to the company the article is about.
The context
Variety selection in chicory rests on the physical measurement of the roots, long recorded by hand on paper: slow, tedious and exposed to human error.
How AI comes in
With delaware and the Digital Wallonia 4.AI programme, COSUCRA set up automated root measurement using computer vision and AI algorithms. Recording becomes faster, more precise and directly usable, drawing on decades of historical data.
What it means for you
Wherever measurements or visual checks are repeated in bulk (quality, sorting, metrology), computer vision saves time and adds reliability. COSUCRA is ultimately aiming at a shorter variety-development cycle; as the published gains remain qualitative, we avoid any unverified figure.
A public company case, restated by Blue Lemon Agent for illustration. No affiliation.
Editorial responsibility. This article is reviewed and approved before publication by the publication director, Ronan Le Boulaire, who bears editorial responsibility for it (see the legal notice). Writing-assistance tools may be used beforehand; no text is published without that human review.
A similar use case in your organisation?
In 15 minutes, let's see how to apply it in your organisation — hosted in France, supervised.