Russian AI Ensures the Perfect Fertilizer Granule Size
The System Monitors and Adjusts Granule Dimensions in Real Time
Researchers at the Advanced Engineering School of Novgorod State University have developed an AI-powered system that uses neural networks and computer vision to control the quality of mineral fertilizers in real time. The technology automatically analyzes granules as they move along production lines, learning to adapt to different manufacturing processes and fertilizer types.
According to Vladislav Rysev, the project’s lead developer, traditional image-recognition methods like Roberts, Sobel, and Canny filters failed to detect the outlines of fertilizer granules — which often appear as a “white blur” due to their small, uniform size.
Beyond Fertilizers
The system is installed above conveyor belts and scans granules using a specialized camera. The neural network instantly recognizes contours, measures the distance to focus, and determines particle size, comparing results to reference data and displaying them in real time. This allows operators to adjust production immediately.
The platform is versatile — it can analyze fine granules in chemical plants as well as larger objects like rock fragments, spherical components, or industrial materials.
The project received a grant from Russia’s Ministry of Science and Higher Education under the Decade of Science and Technology initiative. Researchers say it marks a step toward AI-integrated manufacturing, where smart vision systems ensure consistent quality — not just in fertilizer production but across multiple industries.
