ORAFOL Fresnel Optics is leveraging deep learning to solve a high-precision manufacturing bottleneck that manual inspection could no longer sustain. By deploying Zebra Technologies’ AI machine vision solution, the German-based manufacturer has automated the quality control of over 1.5 million optical lenses. This transition from manual oversight to an intelligent, automated workflow addresses the extreme difficulty of inspecting microstructures on components as small as 4mm. The move signals a shift toward closing the gap between physical production and digital execution layers in high-volume, high-complexity manufacturing environments.
ORAFOL Fresnel Optics Automates Microstructure Inspection
Before implementing the new system, ORAFOL Fresnel Optics relied on manual inspection processes that required operators to hold individual lenses up to specific light conditions to identify defects. This repetitive manual task was prone to human fatigue and necessitated up to 15 operators to support active production lines. The complexity of the Fresnel plastic lenses, which feature intricate microstructures, made finding a standard off-the-shelf inspection provider difficult. To overcome these technical hurdles, ORAFOL required a bespoke approach to handle the unique geometry of their components.
The company ultimately selected a solution powered by Zebra Aurora Vision Studio with Aurora Deep Learning. Because the Aurora software is compatible with any camera, ORAFOL was able to utilize a bespoke optical turret specifically developed to manage the complex microstructures of the lenses. This specialized hardware, combined with Zebra’s deep learning software, has enabled the company to move toward 24/7 workflows. The implementation has successfully cut inspection times in half while providing end-to-end component traceability for every item processed.
Scaling Deep Learning via Zebra Aurora Vision Studio
As production demands increased, ORAFOL Fresnel Optics expanded its automation capabilities by building a second, more advanced system. This secondary system is designed to inspect various optical windows at high volumes and was developed in collaboration with ID Engineering, a Zebra Industrial Automation System Integrator. This expansion utilizes the same Zebra Aurora Vision Studio and deep learning framework, demonstrating the scalability of the software across different product lines and inspection requirements.
The integration of AI into the frontline allows ORAFOL to define rigid specifications within the system, ensuring that any scratch or defect outside of those parameters results in an automatic rejection. According to Rico Traxdorf of ORAFOL Fresnel Optics, this has resulted in significantly higher consistency across the production process. By automating these high-volume tasks, the company is positioning its frontline workers to pivot away from repetitive manual checks and toward more strategic operational responsibilities.
Key Takeaways
- ORAFOL Fresnel Optics has used Zebra Aurora Vision Studio to automate the inspection of over 1.5 million optical lenses.
- The implementation of AI machine vision has reduced inspection times by 50% and enabled 24/7 production workflows.
- The solution utilizes a bespoke optical turret to inspect complex Fresnel plastic lenses that can be as small as 4mm across.
TechInsyte's Take
In our view, the ORAFOL deployment highlights a critical trend in enterprise manufacturing: the necessity of combining specialized hardware with flexible, software-defined AI. The fact that ORAFOL required a bespoke optical turret signals that "general purpose" AI vision often hits a physical ceiling when dealing with micro-scale complexities. However, by using Zebra’s Aurora Vision Studio, ORAFOL avoided vendor lock-in regarding camera hardware, which is a vital strategic advantage for long-term infrastructure resilience. This case suggests that for high-precision sectors, the real value of AI lies not just in the algorithm, but in the ability to integrate deep learning into highly customized, niche physical inspection environments.
Questions & Answers
How does the new AI system improve manufacturing consistency compared to manual inspection?
The system replaces manual, light-based visual checks—which were subject to operator fatigue—with a digital specification. Once a specification is defined in the Aurora Vision Studio, the system automatically rejects any component, such as a lens with a scratch, that falls outside those exact parameters.
What specific technical challenge did the Fresnel lenses present to standard inspection tools?
The lenses feature complex microstructures and can be as small as 4mm across. These physical characteristics made it difficult for standard inspection providers to offer a solution, eventually requiring the development of a bespoke optical turret to work alongside the AI software.
Can the Zebra Aurora Vision Studio software be used with non-proprietary hardware?
Yes. The software is compatible with any camera, allowing ORAFOL to focus its specialized engineering efforts on the optical turret rather than being restricted to a specific camera manufacturer.
How has the implementation affected ORAFOL's labor allocation and production capacity?
The automation has cut inspection times in half and enabled 24/7 workflows. This shift has freed up frontline workers, who previously required up to 15 operators to support production lines, to focus on more strategic tasks.
Source: ZEBRA TECHNOLOGIES
