Researchers have developed a real-time defect detection system for conveyor belts using an optimized YOLOv11 nano model deployed on a Raspberry Pi 5 with a Google Coral Edge TPU. This low-power edge AI solution aims to address the limitations of traditional GPU-reliant models in industrial settings. The system, trained on over 3,000 augmented samples from an operational cement plant and open-source datasets, achieves a precision of 0.742 and a recall of 0.730. It demonstrates stable throughput of 30-40 FPS with low latency, proving effective for predictive maintenance in harsh industrial environments. AI
IMPACT Enables cost-effective, real-time predictive maintenance in industrial settings by bringing advanced computer vision to edge devices.
RANK_REASON This is a research paper detailing a novel application of existing AI models and hardware for a specific industrial problem. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
