Segezha Group has implemented a proprietary computer vision system at its Sokolsky DOK facility in the Vologda region. Developed by the company's subsidiary, Segezha Digital Solutions, the system automates the orientation of wooden boards for CLT panel production.
The technology utilizes industrial cameras with polarization optics and neural network models to scan boards in real time. It processes data on a microcomputer, sending commands to a turner via industrial protocols. The system operates autonomously, requiring human intervention only when recognition confidence is low.
During pilot tests involving over 50,000 boards, the system achieved a 98% recognition accuracy. The solution is integrated as an external module, with setup taking 1–2 days. Data security is maintained through local processing within an isolated industrial network.
The company expects a payback period of approximately 6 months. Economic benefits stem from reduced downtime, lower maintenance costs compared to foreign alternatives, and decreased waste. The system also tracks production metrics to optimize line efficiency.
Segezha Digital Solutions plans to expand the system's capabilities to include linear scanning and defect detection, such as knots and cracks. The company is considering scaling the technology across its other facilities and potentially offering it to external partners in Russia and friendly nations.