DOMO Polymer Solutions, an international leader in the manufacture of technical polymers, needed to ensure that its pellets maintained impeccable quality standards.
A single black spot—the result of thermal degradation or occasional contamination in the extruder—can compromise critical material properties and lead to complaints in high-value industrial applications.
The challenge added an additional layer of complexity: DOMO produces multiple types of plastics, each with its own shades, textures and optical behaviours. The system had to be capable of inspecting correctly without the need for adjustments or changes by the operator, regardless of the material passing through the line.
Until now, inspection was done manually: an operator had to continuously monitor a large vibrating screen, during 24/7 shifts, to detect when burnt pellets began to appear and activate maintenance actions.
This was a tiring, unreliable process which, despite the operator’s experience, could not guarantee the detection of early defects in real time.
The project presented several highly complex challenges. In fact, DOMO had been looking for an automatic solution for a long time, but no supplier had been able to offer a system capable of reliably detecting black spots, especially considering the actual conditions of the line and the large size of the vibrator.
The main technical challenges of the project were:
1. Continuously inspecting very large areas
Each DOMO vibrator is over a metre wide, which required capturing very high-resolution images.
To do this, between 5 and 6 5MP industrial cameras were installed, generating a composite image of up to 30 megapixels analysed in real time. Analysing such large images in real time with deep learning requires efficient models.
2. Detecting extremely small black spots on a changing product
Defects can measure just a few pixels and the material changes depending on production:
The visual environment also presents constant challenges: shadows, glare, external lights, vapours and continuous changes in lighting.
The model had to:
3. Integration with plant traceability
The customer needed detailed information on each production run:
volume inspected, defects detected, evolution over time and correlation with maintenance actions performed.
ID Vision developed an inspection system for DOMO based on OneVision, our industrial machine vision platform with integrated deep learning.
The system allows:
DOMO is a global leader in the production of technical polymer-based solutions for sectors such as automotive, industry, construction and consumer goods.
Its commitment to digitalisation and advanced quality control reinforces its position as a leader in innovation within the chemical sector.
