Recycled Paper Acoustic Panel Production System with YOLOv11-Based Defect Detection

This study developed a semi-automated recycled paper acoustic panel production system with a YOLOv11-based defect detection to improve product consistency and reduce dependence on manual inspection. The produced system consists of a shredding unit, a washing motor for pulping and ingredient mixing, a molding chamber, a drying setup, and a Raspberry Pi for defect detection. Manual operations were limited to paper feeding, panel flipping, and a minimal button intervention. The YOLOv11 model was trained to detect surface defects limited to cracks, deformations, and incorrect perimeter in real time, achieving more than 85% accuracy, 80% precision, 85% recall, a 90% F1- score, and 85% mAP. The prototype successfully produced panels within ±3 mm of the target dimensions and maintained perimeter error rates below 1.5%. Moreover, the average time required to produce one panel was 1 hour. Acoustic evaluation showed that the produced panels achieved a NRC of 0.7 and favorable STL values very close to those of commercially available acoustic panels. These results demonstrate that this system provides an effective and sustainable solution for manufacturing high-quality recycled paper acoustic panels.