Counting small components on large carriers

Challenge

Small metal contains are to be processed into cream containers. For this, metal containers are placed on large, two-dimensional plates to transport them through the anodizing baths provided for this purpose.

The problem: The "disappearance" or falling off of parts while the pinned metal containers "travel" through the anodizing system. Until now, scrap can only be roughly estimated without appropriate technology.

• Low image resolution 
• Reflections and stray light, distorted image display
• No out-of-the-box AI models appropriate for use

Solution

To determine the number of pinned components (or missing components), a camera takes a picture of the entire carrier at the end of the “journey” and counts the pinned metal containers. This shows how many components actually made it through the production line at the end and how many fell off. A previously trained AI analyzes the image and counts the number of pinned components (object recognition)

Benefit

• End-to-end tracking of pinned components
• Increased transparency of the production process