Vision systems inspect juice bottles

Motion Control & Drives

Electric drive, inverter & motor suppliers

Motion Control & Drive suppliers

Position/motion control suppliers

printer friendly version

August 2007 Motion Control & Drives

Original Juice Co. in Australia found itself with a challenge in achieving consistent placement of closures on filled bottles. It wanted closure on all its different bottles to be applied fully as well as straight and not skewed. It also wanted to be sure that tamper-evident bands on the products were not broken. The company decided to install a state-of-the-art vision system that could detect a variety of imperfections, including shoulder strength of bottles. To meet the production requirements, the vision system needed to inspect up to 300 bottles/minute.

The system integrator chose to use the industrial grade Cognex In-Sight 5400 vision sensors, which are the highest-performance models of the In-Sight family of vision sensors. In-Sight sensors have die-cast aluminium IP67 housings and sealed industrial M12 connectors. These attributes are crucial to withstand the wet citric acid environment of the inspection site that is caused by the juices being bottled and the cleaning chemicals used.

Bottles are inspected by two Cognex 5400 vision sensors: one on each side of the bottle

The application runs on an industrial PC mounted in a stainless steel enclosure, which also houses an Ethernet hub, PLC and various power distribution components.

After the bottles are filled and capped, they travel down the conveyor line, where they are inspected by two vision sensors: one on each side of the bottle. As a bottle reaches each vision sensor, the sensor performs two types of inspection:

* A LED backlight provides the sensor with a silhouette image of the bottle. Back lighting provides the maximum contrast between the product and its background, and is ideal for measuring external edges. Cognex In-Sight vision software tools in each sensor analyse the image for defects and determine whether the bottle is flawed or not.

* A pattern of lines is projected onto the bottle and cap and the software determines any deformity in the cap height, cap presence, tamper-band presence and cap skew. To analyse these measurements the software tools use edge detection and histogram measurements.

If either sensor detects a fault, a signal is sent to the PLC, triggering a reject mechanism that removes the product from the line.

Because of the number of variations in the type of bottles, the mechanical fixtures and test procedure selection were designed to accommodate easy product changeover. The cameras and lights are mounted on a fixed bracket that can be moved vertically by a turn wheel to accommodate the different bottle heights. On product changeover, the operator turns the wheel until a reference pointer is in line with the top of the bottle, and then selects the appropriate test procedure by selecting the bottle type on the computer screen.

When the line is running, the last failed image from each vision sensor is displayed on the computer screen. The test result data is recorded in text files on the PC, providing quality records.

Share this article:

Motion Control & Drives

Vision systems inspect juice bottles

Further reading:

Published by Technews