Reproduce “Process Integrated Computer Vision for Real-Time Failure Prediction in Steel Rolling Mill” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Vaibhav Kurrey, Sivakalyan Pujari, Gagan Raj Gupta
Researchers at IIT Bhilai developed a process-integrated computer vision system for real-time failure prediction in steel rolling mills, fusing live visual data from industrial cameras with existing sensor data. The system achieved a 94.2% [email protected] for anomaly detection and reduced false alarms to 2.8%, successfully preventing approximately 10 costly "cobble" incidents per month and yielding estimated savings of Rs. 1.15 Cr.
Key insights
- Effective fusion of real-time machine vision and sensor data provides comprehensive context for anomaly detection, overcoming the limitations of single-modality monitoring systems in complex industrial settings.
- An independent, GPU-accelerated video processing server architecture ensures system scalability and avoids imposing computational overhead on existing, mission-critical industrial process control systems (PLCs).
- Customized deep learning models and careful camera configuration enable robust, high-speed performance in harsh industrial environments characterized by extreme heat, dust, vibrations, and dynamic lighting conditions.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.