Multicamera defect detection system for plastic containers

Overview

The multicamera defect detection system is an innovative solution designed to enhance the quality control process in plastic container manufacturing. By leveraging advanced computer vision and deep learning technologies, this system can detect surface defects such as pearlescence, haze, and crystallinity in real-time. The system captures high-resolution images across the container's surface using a calibrated multicamera setup, providing a comprehensive view that aids in optimizing the blow molding process.

Technical specifications

Key features:

• Multicamera System: High-resolution cameras with overlapping views capture detailed images of various container shapes.
• Deep Learning Models: Utilizes pre-trained models like YOLO and Faster-RCNN for rapid and accurate defect detection.
• 3D Reconstruction: Combines pixelwise classification with traditional computer vision techniques to create 3D visualizations of defects.
• Visualization: Outputs standard mesh file types (e.g., STL, PLY) for easy integration with existing visualization software.

Technology readiness level

Currently, the technology is at TRL 3. Each component—multicamera system design, deep learning model development, and visualization—has been established in other contexts but not yet demonstrated in this specific application of blow molding defect detection.