A Corner-Analysis-Based Method for Dimensional Inspection and Defect Detection of Glass Chamfers

Area of Science:

• Materials Science
• Manufacturing Engineering
• Computer Vision

Background:

• Edge chamfering is crucial for glass strength and safety, yet current inspection methods are inefficient and subjective.
• Automated production lines require faster, more objective quality control for glass chamfers.
• Existing inspection techniques struggle with compatibility with production line speeds.

Purpose of the Study:

• To develop an automated, vision-based inspection method for transparent glass chamfers.
• To improve the accuracy and efficiency of chamfer quality assessment in manufacturing.
• To enable real-time defect detection and dimensional measurement of glass edges.

Main Methods:

• Utilized machine-vision image processing to enhance chamfer features and improve detection.
• Developed a corner analysis technique with an endpoint-screening coordinate system for stable localization.
• Integrated quadrant-based spatial constraints for enhanced endpoint stability and automatic geometric parameter measurement.
• Established defect-discrimination criteria based on geometric tolerances and spatial abnormalities.

Main Results:

• Achieved a mean absolute measurement error of 0.087 mm for dimensional accuracy.
• Recorded an average inspection time of 0.505 seconds per workpiece, demonstrating high efficiency.
• Successfully identified chamfer defects, validating the method's detection capabilities.

Conclusions:

• The proposed vision-based method offers accurate dimensional measurement and effective defect detection for glass chamfers.
• The system is suitable for online inspection in automated glass manufacturing, overcoming limitations of traditional methods.
• This approach enhances operational safety and production efficiency through automated quality control.