Our solution uses an efficient 2-step approach. In the first step, we apply AI-based Vision to segment individual products from higher resolution images of the moving conveyor belt. Here, disturbances such as leaves, sand, twigs or other contaminants are excluded. This ensures a clean dataset of relevant products. In the second step, we use rule-based Vision techniques to measure various characteristics for each product, including length, width, surface area, cylindricity and curvature.
All measured characteristics are stored per individual product in an SQL database. This not only provides structure and traceability, but also makes it possible to generate customer-specific reports based on desired parameters, measurement ranges or time periods.
This technology provides seed breeding companies with a powerful tool for their R&D work. Instead of manual, inconsistent measurements, large volumes of product can now be quickly and objectively analyzed post-harvest.
With this innovation, Vision Partners is contributing to a future where data-driven plant breeding will become the standard. Our expertise in image processing, automation and custom applications comes together in a solution that really works on the shop floor.