AI machine vision halves inspection times

A division of the global ORAFOL Group, German-based ORAFOL Fresnel Optics serves customers worldwide across machine vision, sensor, lighting, display, and solar markets with more than 600 different components. Each item must meet exact quality standards before leaving the facility, including Fresnel plastic lenses which are sometimes as small as 4mm across with complex microstructures, making inspection difficult.

Previously, optics inspection required ORAFOL operators to hold each lens to the light under specific light conditions to find defects. This repetitive task caused fatigue and required up to 15 operators to support production lines. ORAFOL Fresnel Optics needed a solution to automate inspection, but struggled to find a solution provider who could handle the complex Fresnel lenses, so a bespoke solution was needed.

The new quality inspection solution selected by ORAFOL is powered by Zebra Aurora Vision Studio™ with Aurora Deep Learning. It has cut inspection times in half, enabled 24/7 workflows and end-to-end component traceability and freed up frontline workers to focus on more strategic tasks.

“The biggest benefit is that everything is more consistent now,” said Rico Traxdorf, production and technology, ORAFOL Fresnel Optics. “We define a specification in the system and every scratch that is outside the specification is rejected.”

Zebra’s solution allows ORAFOL to save time and money. A standard camera could be used, as the Aurora software is compatible with any camera. A bespoke optical turret had to be developed for the complex microstructures.

“What was initially a challenge that no supplier could solve has now become a key element of the production process for our high-volume products. The system plays a crucial role in ensuring the consistently high, flawless quality that our customers expect from ORAFOL,” said Alexandra Frenzel, quality engineer, ORAFOL Fresnel Optics.

As demand grew, ORAFOL Fresnel Optics built a second, more advanced system to inspect various optical windows at high volume, in collaboration with ID Engineering, a Zebra Industrial Automation System Integrator, again using Zebra Aurora Vision Studio with deep learning.

www.zebra.com