Client and Challenge

A manufacturer of high-end ceramic components relied on manual visual inspection. This process was subjective and fatiguing, resulting in a 7% "escape rate." Defective products reaching premium international clients led to expensive warranty claims, high return logistics costs, and significant damage to their "Made in Germany" brand reputation.

Solution

We developed a bespoke Computer Vision solution deployed on edge devices directly on the conveyor belt. Using high-resolution cameras and a Convolutional Neural Network (CNN), the system identifies microscopic cracks and glaze imperfections in real-time, instantly triggering a pneumatic sorter to remove defective units.

Outcomes

• Warranty & Liability Savings: Slashed annual warranty claims and replacement costs by €310,000 by reducing the defect escape rate from 7% to under 0.5%.

• Significant Revenue Upside: The 40% increase in inspection throughput allowed the plant to clear production bottlenecks, enabling the fulfillment of an additional €1.5M in annual orders without increasing headcount.

• Operational Margin Improvement: Reduced the cost-per-inspection by 60% and eliminated the financial waste of shipping "scrap" material to overseas distributors.