Wine labeller

412 views

Published on

0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
412
On SlideShare
0
From Embeds
0
Number of Embeds
10
Actions
Shares
0
Downloads
0
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Wine labeller

  1. 1. A Vision system for a labelling machine Student: Alberto Caccia Thesis presentation, November 2010
  2. 2. The topic: labelling (good) wine bottles Front label DOCG label Back label Requested byCantina Leonardo da Vinci Developed by
  3. 3. Labelling machine <ul><li>Manual operations </li></ul><ul><li>Wrong labels on bottles </li></ul>Three labels/sec
  4. 4. Hardware <ul><li>Smart Camera from National Instrument </li></ul><ul><li>Full operation integrated into the smart camera </li></ul><ul><li>Real Time operation system </li></ul><ul><li>LabVIEW Real Time </li></ul><ul><li>time constraint: 360ms </li></ul>Illumination PLC of the labelling machine Detect the bottle Image acquisition & elaboration 3 reject trigger 1 2 strobe
  5. 5. Elaboration of DOCG labels <ul><li>Code </li></ul>Denominazione OCR Pattern matching AAY0046054
  6. 6. Enhancement of the OCR performance-1 Alignment correction Italics correction Background correction Character segmentation Thresholding Morphol. analysis Line fitting rotation
  7. 7. Enhancement of the OCR performance-2 Alignment correction Italics correction Background correction Character segmentation Custom algorithm Pixel-based stretching
  8. 8. Enhancement of the OCR performance-3 Alignment correction Italics correction Background correction Character segmentation histogram Custom analysis Gray scale mapping
  9. 9. Enhancement of the OCR performance-4 Alignment correction Italics correction Background correction Character segmentation Detection of letters and numbers Detection of spaces between characters OCR algorithm ABA06847006
  10. 10. Elaboration of front and back labels-1 Many different labels share three details
  11. 11. Elaboration of front and back labels-2 <ul><li>Template matching used to detect the details </li></ul>Detail 1 Detail 3 Detail 2 Template matching
  12. 12. Flexibility to variation of the labels <ul><li>A suitable panel allows the operator to create the layout corresponding with each new label </li></ul>
  13. 13. Test interface <ul><li>Full, remote control of the camera </li></ul><ul><li>Powerful functions increase the flexibility of the whole system </li></ul><ul><li>Very efficient for: </li></ul><ul><li>data monitoring </li></ul><ul><li>File exchange </li></ul><ul><li>OCR data file customization </li></ul>
  14. 14. IN-field testing <ul><li>Data analysis on the number of rejected bottles during the application of correct labels </li></ul><ul><li>Rejection due to </li></ul><ul><li>Unrecognized labels Pattern Matching failed </li></ul><ul><li>Invalid DOCG codes OCR failed </li></ul><ul><li>Unreadable DOCG codes No characters in the set </li></ul>Inspected bottles: 8550 Unreadable codes: 1 Rejects: 404 ( 5,5% ) OCR errors: 63 ( 0,7% ) Inspected bottles : 2290 Unreadable codes: 7 Rejects: 35 ( 2,5% ) OCR errors: 15 ( 0,98% ) Inspected bottles: 2580 Unreadable codes: 0 Rejects: 66 ( 2,6% ) OCR errors: 1 ( <0,1% )
  15. 15. Conclusions <ul><li>A Real Time vision system using a Smart Camera monitors a labelling machine. </li></ul><ul><li>LabVIEW e LabVIEW Real Time </li></ul><ul><li>Suitable Pattern Matching and OCR algorithms </li></ul><ul><li>OCR optimization </li></ul><ul><li>Good system flexibility and efficiency </li></ul>In-filed testing <ul><li>DOCG labels recognition software </li></ul><ul><li>Front & back labels: the test is IN PROGRESS </li></ul>Future work <ul><li>Automatic control of the camera, PC-based </li></ul><ul><li>Integration to the supervisor system </li></ul>

×