Ai-based defect detection in molded containers using light scattering analysis

Overview

This cutting-edge AI technology enables the detection and localization of defects in molded containers by analyzing light scattering patterns and temperature profiles. Utilizing a high-speed thermal camera and a specialized light source, this solution captures the necessary data to automatically identify defects such as heat haze and overstretch-pearlescence. By correlating these defects with the 3D temperature profile of preforms, manufacturers can make informed decisions to optimize the heating and conveying processes, significantly enhancing quality control.

Technical specifications

• Defect detection: Utilizes a specially selected light source and camera to capture light scattering patterns indicative of defects.
• Temperature profiling: Employs a high-speed thermal camera to acquire 3D temperature profiles of preforms prior to molding.
• Machine learning: Applies semantic segmentation for classifying defect types as "heat haze", "overstretch-pearlescence", or "no-defects".
• Robust analysis: Capable of maintaining accuracy despite variations in ambient lighting conditions.
• Correlation analysis: Links defect positions and types to temperature profiles for corrective actions in the manufacturing process.

Technology readiness level

This technology is currently at TRL 2, indicating that it is in the early stages of development, with basic principles observed. Future validation plans include further training and evaluation of machine learning models, assessing the optimal illumination setup, and correlating thermal and defect data for improved detection accuracy.