PETKUS OS P — AI-powered optical sorting to predict quality and turn losses into value

The OS P combines high-resolution imaging with Seed-X AI to sort seeds by predicted outcome (e.g., germination, usable plant, genetic purity). Each seed is scanned and characterised across countless visual parameters; after a one-time grow-out training the system builds a classifier that links image features to real field/greenhouse results. Operators can then choose the optimal balance between quality (germ %) and yield (reject rate), converting low- or even “zero-value” lots into sellable product while meeting market specs. This data-driven approach improves planning and inventory utilisation—helping ensure that every plot contributes saleable volume, with the potential to recover the 10–15% of annual lots that are typically unsellable after conventional processing.

Key advantages

• Sort by germination / usable plant: first optical sorter dedicated to outcome-based seed selection.

• Seed-X classifier: AI built from imaged seeds linked to grow-out results; once trained, sorting is automatic and reproducible.

• Profit optimisation: set target germination and allowable rejects to maximise margin and sustainability.

• Multiple traits in one platform: supports germination, uniformity, ploidy, genetic purity, blind seeds, deformation, skin damage, and debris removal.

• Simple operation: extra-large touch panel; classifier and thresholds managed through an intuitive UI.

• Networked support: Windows-based platform with remote connectivity for updates and assistance.

How it works

1. Image & place each seed in trays.

2. Grow-out test in greenhouse.

3. Evaluate usable plants.

4. Aggregate data in OS P.

5. Train classifier with Seed-X AI.

6. Sort seeds to the specified use case and quality target.

Typical applications

• Germination uplift, usable-plant selection, uniformity improvement.

• Differentiation of ploidy level and genetic purity.

• Elimination of blind/deformed seeds, skin damage, and plant debris across multiple crops.