Ceramic Disc Machine Vision Retrofit

Impetus to Invent

When you’re producing 90 medical parts a minute and can’t afford a single defect, you need machine vision. The current manual sampling plan just wasn’t cutting it, and something had to be done. Enter Invent Automation.
Implemented between the cleaning line and the downstream equipment was our custom bowl feeder that fed parts from the washroom into the clean room via conveyor. From here, our deep learning machine vision algorithms took over and inspected every part, passing or failing as they went by. If a part was bad, it was kicked out in real-time. If good, parts were separated into two lanes and passed to the downstream equipment. Additional sensors validated that if a bad part was found, it was confirmed as having been sent to the reject bin, a critical step in medical applications. Now the customer could rest assured that every part was being inspected every time.

Problem Solved

Replacement of manual sampling inspection with high-speed machine vision.

Lessons Learned

UHMW plastic wears the least against nasty ceramic dust, even better than steel!

What Was Unique

Feeding, inspecting, sorting and separating in such a tight space was a lot of fun!

What Was Challenging

In a tight space separated by a wall we had to integrate a bowl feeder, two conveyors and a machine vision inspection system. Additionally, we had to integrate with the upstream and downstream controls systems.