A Very High-Speed Validation Scheme Based on Template Matching for Segmented Character Expiration Codes on Beverage...

Area of Science:

• Industrial Automation
• Computer Vision
• Quality Control

Background:

• Previous work introduced MONICOD, an industrial validator for can expiration codes.
• MONICOD operates at high speeds (200 frames/sec, 35 cans/sec).

Purpose of the Study:

• To describe the final character validation stage of the MONICOD system.
• To introduce learning algorithms for building morphology databases.
• To evaluate the performance of the character validation process.

Main Methods:

• Character validation using a template matching scheme based on bitwise operations ('morphologies').
• Development of two learning algorithms for creating valid morphology databases.
• Comparison of acquired character shapes with expected shapes.

Main Results:

• A single false positive (0.21%) occurred in validating 465 cans.
• Validation time is at least 20% faster compared to traditional classifiers.
• Achieved error rates comparable to Support Vector Machines (SVM) and k-Nearest Neighbors (KNN).

Conclusions:

• The described character validation method enhances MONICOD's accuracy and speed.
• The system offers a competitive alternative to existing machine learning classifiers for expiration code validation.
• MONICOD provides an efficient and reliable solution for industrial can inspection.