A Character Segmentation Proposal for High-Speed Visual Monitoring of Expiration Codes on Beverage Cans

Area of Science:

• Food science and technology
• Industrial automation
• Computer vision

Background:

• Expiration date labels are essential for food safety and consumer information.
• Accurate reading of expiration codes on food packaging is critical for quality control.
• Industrial systems face challenges with high-speed printing and reflective surfaces.

Purpose of the Study:

• To develop an improved method for reading expiration codes on food packaging.
• To enhance the MONICOD (MONItoring of CODes) system's performance in industrial settings.
• To address challenges posed by high printing speeds and metallic surfaces.

Main Methods:

• Implementing a novel shape extraction technique for code reading.
• Utilizing advanced image processing to overcome lighting issues on metallic surfaces.
• Integrating the solution into the MONICOD system for real-time validation.

Main Results:

• The new method significantly improves the speed and accuracy of expiration code reading.
• Quantitative results demonstrate enhanced performance in extracting shapes from challenging surfaces.
• The system successfully handles high-speed production lines (35 cans per second).

Conclusions:

• The developed solution effectively enhances the MONICOD system for industrial expiration code validation.
• This advancement contributes to improved food safety and quality control in manufacturing.
• The method offers a robust approach for machine vision in challenging industrial environments.