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

作者：Rodriguez Rodriguez Jose C; Quesada Arencibia Alexis^*; Moreno Diaz Roberto Jr; Garcia Carmelo R

来源：Sensors, 2016, 16(4): 527.

DOI：10.3390/s16040527

• image processing
• OCR (Optical Character Recognition)
• pattern recognition
• industrial inspection
• very high-speed computing
• character segmentation
• tracking

Expiration date labels are ubiquitous in the food industry. With the passage of time, almost any food becomes unhealthy, even when well preserved. The expiration date is estimated based on the type and manufacture/packaging time of that particular food unit. This date is then printed on the container so it is available to the end user at the time of consumption. MONICOD (MONItoring of CODes); an industrial validator of expiration codes; allows the expiration code printed on a drink can to be read. This verification occurs immediately after printing. MONICOD faces difficulties due to the high printing rate (35 cans per second) and problematic lighting caused by the metallic surface on which the code is printed. This article describes a solution that allows MONICOD to extract shapes and presents quantitative results for the speed and quality.

• 出版日期2016-4