🔍 Seeing the Invisible: Real-Time Defect Detection on Painted Aluminum Plates

In industries where surface perfection is mandatory — from automotive to aerospace — even the tiniest scratch can trigger rework, delays, or warranty claims. Yet, manual inspection is slow, subjective, and costly. In May 2025 at SPS Parma, we set out to show how this challenge can be solved — not with theory, but with tech you can put into production today. We unveiled a fully operational demo line, inspecting paint-finished aluminum plates, using our NEO machine vision system, designed to catch even the most subtle scratches, dents, or surface anomalies in real time.

‍

And this wasn’t just a showpiece. This setup mirrored what our customers rely on 24/7 in their actual production lines— ensuring every single product meets strict visual quality standards.

‍

🤖 The Setup: Smart Motion Meets Precision Vision

Our demo system at SPS Parma brought together collaborative robotics, 3D vision, and high-performance imaging — all orchestrated to replicate a real-world, fully autonomous inspection cell.

‍

⚙️ Motion & Part Handling: Collaborative and Flexible

At the heart of the system was a collaborative robot equipped with a 3D vision camera, enabling safe and flexible operation in a dynamic environment.

• The 3D vision system allowed the robot to locate and pick parts from an unsorted tray — eliminating the need for exact positioning or fixturing.
• Once a part was detected, the robot used compressed air and vacuum suction to gently lift it.
• It then precisely unloaded the part into the imaging station — ensuring exact alignment under a calibrated inspection setup.
  

This hands-free handling ensured both speed and repeatability, even with loosely arranged components.

‍

🔍 Vision & AI: Precision Optics + Edge Intelligence

The vision system was built for a single goal: detecting subtle surface anomalies at production speed — consistently, reliably, and without manual intervention.

‍

🔬 Optical Setup: Built for Surface Perfection

High-performance inspection starts with the right optics. Thanks to our new application lab — developed in collaboration with Opto Engineering — we tested and selected the ideal imaging configuration for this challenge: a telecentric lens combined with a coaxial illumination.

• Telecentric optics eliminate perspective distortion, ensuring accurate dimensional analysis and consistent imaging regardless of part placement.
• Coaxial lighting delivers uniform illumination across the entire field of view — essential for detecting surface-level defects like scratches, dents, or coating inconsistencies on reflective materials like painted aluminum.
• The result: sharp, uniform images with minimal glare, shadowing, or false positives.

Want to know why this setup outperforms traditional lighting in surface inspection? Contact our Trinity Team — we’d love to explain.

🧠 NEO Software: Real-Time AI on the Edge

Once the image is captured, analysis happens instantly.

Our proprietary NEO inspection software runs directly on a compact and cost-effective NVIDIA Jetson IPC — eliminating the need for cloud computing or high-end servers.

• Inference time: ~7 milliseconds per frame
• Fully offline, on-device processing
• No latency, no network dependency — just real-time, production-grade AI

This makes NEO a perfect fit for edge deployments in real manufacturing environments, where reliability and speed are critical.

‍

‍

🧠 Instant Feedback with Real-World Interaction

Here’s where it got fun.

Visitors were invited to intentionally damage some of the otherwise flawless aluminum plates — with carpet knives, pens, or whatever tools were at hand. Then, we ran them through the system again.

Result?

Within seconds, our AI picked up the newly created micro-defects and visualized them with a color-coded heatmap, instantly appearing on our custom dashboard.

This live interaction made it crystal clear:

• You can trust machine vision to catch what the human eye misses.
• You can interact with inspection data — zoom, track trends, flag patterns.
• You can integrate it into a real production loop, not just a lab setup.

‍

💡 Why It Matters

In industries like automotive, luxury, and many others, surface perfection isn't optional — it's expected.

With solutions like NEO, we’re helping manufacturers:

• Reduce defect rates and rework
• Gain traceability over every part
• Enable smarter quality loops with visual, interactive data

And we're just getting started.

‍

Want to see how it looked in action?
Check out our live footage from SPS Parma 👇

‍

‍

👋 Let’s Talk

If you’re working with materials where surface perfection matters — from polished metals to coated plastics — we’d love to show you how this tech can scale to your process.

‍