At a glance
AI image recognition replaced the existing reader to inspect lot and expiry printing on every foil pouch inline, across light and dark packaging, triggering existing machine signals on failures with no re-tuning after installation.
The problem
The line packs products in foil pouches of different materials and colours. After printing, the existing system could not reliably read the print, leading to missed and false rejects. Even with manual sampling or a second check, defects still slipped through and caused customer complaints, and the reader had to be re-tuned by hand before every run.
Project goals
- Replace the existing reader with AI image recognition for 100% inline inspection
- Handle pouches of every colour
- Trigger a signal on print failures that works with existing equipment
- No re-tuning after installation, with optional lot and date verification
Light and dark packaging
Features
- Dynamic localisation: captures a wider image, then finds and crops the print area automatically so print position variation does not affect recognition.
- AI print reading: an AI object detection model replaces classic OCR and adapts better to printed characters, greatly reducing false rejects without missing defects.
- Image records: every reject keeps a snapshot, and a random sampling ratio can be set per lot for later review.
Benefits
- From manual sampling to 100% inline inspection, reducing escapes and complaints
- Light and dark packaging both recognised, with no re-tuning at changeover
- Signals connect to existing equipment without replacing the line
- Every reject is backed by an image record for traceability
