Quality Control & Process Monitoring
Computer vision systems, laser scanners, and digital twins enable real-time anomaly detection and adaptive control, closing the feedback loop between sensing and actuation for consistent part quality.
Hassan et al., Composites Part C (2024) - Section 5 & Quality Control Focus
Key Impact Metrics
0.44/0.69
Precision/Recall for SSD+VGG16 defect detection (IoU 0.4)
86.5%
Accuracy for AI-based surface inspection
90%
CNN anomaly classifier accuracy with laser point-cloud
Highlighted Studies
| Study | Highlights |
|---|---|
|
SSD + VGG16 Pipeline
Real-Time Object Detection
|
Trained on 2,500 annotated frames, reached Precision/Recall up to 0.44/0.69 (IoU 0.4) before pausing faulty jobs via Raspberry Pi controllers. |
|
AI Surface Inspection
Pre-processing + Shape Analysis
|
Classified extrusion defects with ~86.5% accuracy, with findings pointing to lighting-tuned deployments for production environments. |
|
Laser Point-Cloud Monitoring
CNN + PID Control Loop
|
Paired with CNN anomaly classifiers achieving 90% accuracy, fed PID loops that maintained layer height within 0.1% deviation. |
Quality control systems are most effective when integrated with process optimization—detected defects trigger parameter adjustments in real time.