Reliable detection, intelligent verification
3 June 2026
What if camera-based Optical Character Verification (OCV) could check alphanumeric markings, such as the best-before date (BBD), and assess them correctly even when the quality of the inkjet coding deteriorates? The modular HEUFT system's versatile AI cameras and deep-learning-enabled image processing make it possible!
Last autumn at drinktec, we had integrated our HEUFT reflexx A.I. camera and the image processing of the same name with new deep-learning OCV into the HEUFT canLine II for high-speed empty can inspection. At interpack 2026, our in-house developments for the intelligent verification of alphanumeric inkjet codes were recently demonstrated in the HEUFT PRIME full container check, which, using lateral profile measurement, can now also examine the tightness of freshly filled beverage cartons. The camera and image processing also enable intelligent verification of best-before dates, batch and lot numbers on the packaging – and reliably identify incorrect information regarding the nature and shelf life of the product contained within.
To ensure this still works even when the print image is compromised because the print quality of the inkjet coder deteriorates over time, countless relevant images have been trained into a dedicated deep learning model. HEUFT reflexx A.I. uses it to separate and centre each individual character. The intelligent image processing assesses different variations of recognised letters and numbers more precisely and competently than ever before.
Fewer stops, fewer false rejections
Despite printing errors and display issues such as misaligned ink dots, distorted, smudged or blurred letters, numbers and lines of text, reliable content verification is achieved even at high line 
speeds of 1,200 packs per minute and above. The inkjet printer does not need to be repaired immediately, nor does the relevant filling or packaging line need to be halted specifically for this purpose. To ensure efficiency and productivity, printer maintenance can instead be postponed until regular line stops. Excessively high false rejection rates due to declining coding quality are avoided, and valuable packaging and foodstuffs are not wasted.
Whether during empty can inspection or only when checking full containers: Like all technologies from our modular kit, the HEUFT reflexx A.I. camera and AI-based image processing can be flexibly integrated into virtually any of our systems to implement, amongst other things, a deep learning OCV. Even in challenging situations, they maintain the high efficiency of entire filling and packaging lines.
