Kordiant holding has implemented a computer vision system at its Gislaved plant in Kaluga to automate quality control on the rubber mixing line. Developed by the company's internal engineering team, the solution replaces manual visual monitoring with an autonomous system based on a USB camera and an industrial computer.
The system utilizes a local machine learning model that processes images at a rate of up to 10 frames per second. It identifies the presence of chemical component packages on the conveyor and detects potential jams by monitoring the movement cycle. If a blockage occurs, the system sends an immediate signal to the line controller to manage the mechanical components.
The ML model was trained on a dataset of video footage captured directly from the production floor. The software maintains a complete history of operations, including video recordings of each cycle, which are accessible via a dedicated web interface. The system operates autonomously without relying on external services.
The implementation has eliminated the need for constant manual oversight and prevented unplanned downtime on the rubber mixing line. Kordiant is expanding its use of computer vision technology, having previously installed cameras capable of recognizing over 30 tire sizes at its Yaroslavl tire plant.
According to Kordiant CEO Dmitry Gorbachev, the company is focused on automating quality control across all production stages to improve process transparency and minimize losses. These initiatives are part of a broader strategy to establish new digital quality management standards within the Russian tire market.
