Dpoint: Computer Vision Quality Control
The Challenge
Dpoint, a precision components operation in the Midlands, was struggling with an inconsistent quality control process. Manual inspection was catching roughly 80% of defects, but 3% of shipments were still triggering customer complaints from tier-one automotive clients. Recruiting skilled inspectors was increasingly difficult, and the quality manager was concerned about audit exposure against AS9100 expectations.
The Solution
We deployed a bespoke computer vision quality control system on two production lines, trained on a labelled dataset of historical defect images and integrated with the factory's existing MES. The vision models flag defects in real time, route borderline cases to human inspectors with annotated images, and maintain a fully auditable record of every inspected part, which has materially improved Dpoint's audit readiness.
Results
The Numbers Speak
Defect Detection Rate
Annual Cost Saved
Customer Complaint Rate
Faster Line Inspection
“Our tier-one customers have noticed the difference. Complaints are down, audit evidence is airtight, and our inspectors can focus on the genuinely tricky parts rather than repetitive checks. It has changed how we compete.”
Shuaib
Chief Technology Officer, Dpoint