Quality Magazine

Case study: High-Speed Bottle Inspection

May 1, 2007
Control of swing cap bottles and labels must be done at high speeds. Source: MVTec


German beverage industry specialist BBull (Königsbach-Stein, Germany) faced a challenge to inspect tens of thousands of containers an hour. BBull’s high-speed machine vision system Stratec BV 3000W effects an all-around control of labels, tins and bottles.

The system must control labels on the front, back and at the bottle’s neck. It checks the label’s fit, reads the expiration date, appraises the quality of captions, detects scratches, controls graphic elements as logos and discovers dog-ears. Bar codes and data matrix codes must be identified and read neutrally against rotation and position.

All these features are required by today’s 100% quality control in industry. BBull is a specialist for quality control in bottling and tinned food, with 15 years of machine vision experience in these industries.

The Halcon-based machine vision software is installed at the Baileys plant in Ireland. Source: MVTec

To handle enormous quantities of test items, machine vision software is needed that not only shows reliability and robustness, but also is able to manage an enormous amount of data at the highest speeds. Up to a hundred thousand inspections an hour are possible.

BBull’s engineers settled on the machine vision software library Halcon from MVTec (Munich, Germany) because Halcon offers automatic operator parallelization (AOP). The Halcon-based software currently runs on an industrial PC under Windows, and the pulsing of the devices is about 1.8 gigahertz. The fitting of hardware and Halcon’s effectiveness was sufficient to perform the postulated requirements of the system.

“Without Halcon, our project with this high performance would have been impossible,” says Georg Krauss, chief technology officer for control systems at BBull’s. “Halcon is able to rule this speed, to thread this amount of data and to reliably process them.”

Cameras and illumination devices are key components to a machine vision system. Source: MVTec

To operate this enormous amount of containers, the machine vision application must acquire and process four camera images in parallel running on a single PC.

Furthermore, the system supports other complex tasks. BBull has developed a procedure to control swing cap bottles. The metal bails must be in the correct position, cap and packing ring adjusted, the seal label undamaged, and defilements or damages in the aperture area must be discovered. Moreover, the barrels must be checked for completeness before the packing process. The Halcon-based machine vision software can be addressed by a touch screen, allowing all relevant data to be accessed and easily revised. To integrate the image processing management system into existing industrial bottling systems, BBull offers the necessary interfaces, therefore, redundancies can be avoided and workflows can be eased.



MVTec Software GmbH
+49 89 457695-0
www.mvtec.com
Reply 20


sidebar: benefits

  • The Halcon-based machine vision software can be addressed by a touch screen allowing relevant data to be accessed and easily revised.
  • To integrate the image processing management system into existing industrial bottling systems, BBull offers the necessary interfaces to avoid redundancies and ease workflows.
  • The Halcon-based software offers automatic operator parallelization (AOP), and currently runs on an industrial PC under Windows.