Managing the Demand

Rexona deodorant is one of the most popular products manufactured by Unilever in its Philippines facility.  Due to the increase in production, it became an obvious need for Unilever to automate its manufacturing process.

The Decision Process

Prior to implementing a machine vision system, Unilever was engaging up to 10 operators per line to perform the following quality checks manually:

• Presence and integrity of the cap
• Correct color and date code printed on the cap
• Labels inspection on both sides of the product

The production lines run at an average of 220 bottles per minute.  Despite investing heavily in a well-trained labor force, Unilever faced a high volume of rejects because the labels were either poorly aligned or wrinkled during the application process.  Performing manual inspection on these tasks had become increasingly unrealistic.  Furthermore, manual inspection did not provide a check of the printed barcodes on the packages.

Gilbert Plumos, Product Line Engineer for Unilever, decided to improve production performance by exploring various machine vision systems.  He understood the need to have a vision system in place which would allow operators to be redeployed to other complex and mission critical tasks on the production floor.

The Evaluation Process

After a few rounds of collecting the inspection criteria, Plumos finally decided that a three-camera inspection system would best suit the plant’s needs. A color smart camera would be positioned overhead to check for cap presence and integrity, cap color and date code validation; a second color camera would be positioned at the front of the product to check label alignment and verify the color of product label; and a third camera would look at the rear side of the package to perform barcode reading and verify the product code.

“We put Cognex, and other leading machine vision suppliers to the test during the evaluation process.  I must say I am most pleased with the results that came from the In-Sight vision systems, especially its OCR performance and color recognition ability,” says Plumos. “And the In-Sight Explorer software comes with the EasyBuilder interface which walks the user through a simple step-by-step process of setting up a vision application. Our guys love it!”

The Integration Process

The greatest challenge was to select lighting that would provide uniform illumination for all the three cameras because the surface of the Rexona bottle is reflective, Furthermore, Unilever expects production yield per hour to grow with an increase of line speed in the foreseeable future.  The lighting was an important element.

Inzpect Technologies, an integrator in the Philippines, built custom lighting and a graphical user interface for Unilever to allow the operator to manage and view the images of all three cameras simultaneously with additional onscreen buttons to control other electrical and mechanical components within the vision module.

This was done seamlessly because the In-Sight vision systems offer a Software Development Kit (SDK), allowing users to create highly customizable, stand-alone user interfaces tailored to specific application needs.

The Result

After the successful implementation of the first vision system providing a fully automated manufacturing line, Plumos is now replicating this success to two other lines. He believes that other divisions within the organization would soon adopt machine vision in time to come.