1. 0 Machine Vision Solutions for Part Identification & Traceability 2D Data Matrix OCR/OCV Linear Barcode Welcome to Part Identification

2. 1 Agenda

3. 2 Agenda Con’t.

4. 3

5. 4 Part Identification & Traceability Overview Dr. Michael Schreiber DVT Director of Applied Engineering

6. 5 Topic Outline Identification and Traceability Methods of Identification OCR 2D DataMatrix Marking Methods Readability

7. 6 Identification The ability to recognize a part by unique key features and/or markings –Feature Identification –OCR –1D Barcodes –2D Codes

8. 7 Feature Identification Size Shape Color Unique Features

9. 8 Feature Identification May be able to use existing features Less Specific Usually go/no go Not unique Less data for traceability

10. 9 Traceability Ability to store pertain data Specific to part, date, lot Data is Valuable –Inventory control, forecasting, business operation –Recall information –Error Proofing – Making sure part are made in the correct order

11. 10 Methods of Reading AlphaNumeric –Ease for Human to read –Takes up a large amount of room –Hardest for Machines to read 1D Codes –Simple readers –No Error Correction –Large Footprint

12. 11 Methods of Reading 1D Stacked Codes –More information, 2D Codes –Small Footprint –Error Correction Build-In –More information

13. 12 Industrial OCR What is Industrial OCR? –Reading of characters in any manufacturing or production environment –Lot Codes, Date Codes, Part Numbers –Typically use OCR “friendly” fonts OCR A, OCR B, Semi Undecorated, non-serif fonts Almost always upper case

14. 13 Industrial OCR Semi-conductor –Wafer ID’s Virgin & in-process –IC part numbers Laser etched –Fixture ID’s –IC pack numbers

15. 14 Industrial OCR Automotive –Metal parts Engine block, cases, stamped body parts, etc. –Glass –Displays Aerospace –Similar to automotive

16. 15 Industrial OCR –General Manufacturing Metal parts –Pin stamped –Engraved –Warehousing Labels –Ink-jet, dot matrix –“Junk mail” sorting

17. 16 Industrial OCR Pharmaceutical –OCR combined with OCV –Date, lot code inspection & sorting –Stringent FDA standards Label printers –OCR and/or OCV –High speed

18. 17 Industrial OCR Cellular Phone –“Read” displays –Check for missing segments Medical manufacturing Many, many more …..

19. 18 OCR Technologies Correlation –Mathematical, oldest, slow, clean images only Geometric –Pattern matching algorithms Algorithmic –Pre-trained, only perfect character printing Feature extraction –OCR specific, fast, flexible, trainable

20. 19 OCR – Feature Extraction Foreign Languages Curved Printing Vertical Printing Dot Matrix Handles more Variability

21. 20 OCR Font Choices OCR enhanced - best Serif or decorated fonts - worst Stroke - best Segmented – problem for most readers, not DVT Dot Matrix – big problem for most, not DVT

22. 21 Typically Tough Font Situations Non-stroke characters Imperfect stroke characters Warped and italicized characters –Dot matrix in this class impossible for most Touching characters Varying size and aspect ratio “Junk” above and/or below characters Imperfect lighting

23. 22 Achieving OCR Success Dot matrix or Stroke fonts –Must be OCR “friendly” Can be thresholded –Not embedded in patterns or pictures –No gross imperfections running through the text REPEATABLE IMAGES!!!!!!!!!! Tougher images can be read, but require more effort

24. 23 Application Overview OCR Smart Sensor is trainable Variety of automatic threshold options OCR specific filters Dot matrix or stroke characters Pre-trained standard fonts Control of spacing and bounding boxes

25. 24 1D Codes UPC/EAN Interleaved 2 of 5 USS-128 USS-39 Codabar PharmaCode BC412 PDF417 RSS-14 Code 93 POSTNET Micro PDF417 RSS –Limited, Composite, Expanded Planet Code UPC Composite

26. 25 2D Data Matrix More Data –6 to 3116 digits Smaller Footprint Reed-Solomon error correction –Up to 50% code degradation More Tolerance to lighting and marking changes Many formats: –Square and rectangular –ECC 00, 050, 080,100,140, 200 –ECC-200 preferred on new applications

27. 26 2D Code Selection Industry Specific –DataMatrix Prevalent Automotive Industry Action Group (AIAG) Department of Defense Aerospace (IAQG) NASA Electronics Industry Association (EIA) –ECC200 for non-specified standard

28. 27 Square 2D Symbols NumbersCharactersBytes 10x10631 12x121063 14x1416106 16x16241610 18x18362516 20x20443120 22x22604328 24x24725234 26x26886442 32x321249160 36x3617212784 40x40228169112 44x44288214142 Larger Codes supported as well

29. 28 Rectangular 2D Symbols NumbersCharactersBytes 8x181063 8x3220138 12x26322214 12x36443120 16x36644630

30. 29 Less Sensitive to Variability Skew Lighting Size

31. 30 Size and Focus

32. 31 Marking Factors Life Expectancy Material to be printed on Surface characteristics Volume Symbol Size Space

33. 32 Marking Methods Laser Marking –Laser burns surface –Good speed and quality –Semicon, electronics, packaging, medical Dot Peening –Mechnical stylus dented surface to create dots –Automotive, Aerospace Ink Jet Printing –Ink leaves colors mark on surface –Surface characteristics determine permanence –High Speed, moving parts, good contrast Thermal Transfer Etching (Electro-Chemical ) –Oxidation of metal surface –Round parts

34. 33 2D Codes

35. 34 Mark Placement Flat Surface Clean Area Raised surface instead of embedded Readily visible

36. 35 Readability/Print Grading Quiet Zone –Area around Code Finder Bars –L Shaped used for finding symbol “Clocking” Pattern –Alternating dots opposite Finder Bars Data Cells –Encoded Information

37. 36 Cells Marking Method determines cell shape and size Minimum size may be determined by surface texture

38. 37 Verification/Grading AIM Standard – A-F Grading –Symbol Decode –Contrast –Print Growth –Axial Non-Uniformity –Unused Error Correction Other Parameters –Dot Size –Dot Offset

39. 38 Reading Challenges Poor Focus Washed Out Low Contrast Non Uniform Background

40. 39 Reading Challenges: Printing

41. 40 Reading Challenges: Background

42. 41 DVT DMReader DataMatrix/Readers

43. 42 Solutions Intelligent Scanners –OCR, 1D, 2D Reading –DMReader frontend for simple setup –Regular Speed and High Speed SmartImage Sensors –Color, Higher Resolution, Stainless Steel Options

44. 43 4 to 8 times faster than the Legend 540 100 DataMatrix reads per second Hardware Specs: 640 X 480 Grey- scale resolution Hatachi SH4 Processor 32 Mb RAM 8 Mb Flash Stainless steel enclosure that is suitable for FDA-regulated wash- down environment IP66,IP67, and IP68 Rated Industrial I/O and Ethernet connectors Legend 550 Speed

45. 44 CFR 21 Part 11 Pertains to –Electronic record keeping –Electronic signatures In Terms of OnLine inspection Systems –Documenting a trail of changes and –Who made the changes

46. 45 CFR 21 Part 11 Two DVT Component allow a system to be configured for FDA compliancy. –Access Control –DVT SmartLogger Service

47. 46 CFR 21 Part 11 Access Control and SmartLogger –Aid in creating a FDA 21 Part 11 compliant process –Increase ability to monitor user activities –Maintain version history of camera states –Provide central location for event logging for multiple systems

48. 47 DVT Support Model Bob Settle DVT Director of Marketing

49. 48 Siemens VDO Bob Segraves DVT Director of Business Development

50. 49 Siemens VDO

51. 50 R&D Tool by Frank Long, DVT Director of Business Development

52. 51 R&D Tool

53. 52 Vision System Wiring Diagram

54. 53 SmartLink & Touchscreen Operator Interface

55. 54 Free Breakout Board!

56. 55 * Free Breakout Board! $300 Value! Must be purchased with camera, lens, & light in next 60 days Code #: XK17BJ Email: monica.berry@dvtsensors.com

57. 56 DVT’s core value system is based on the concept of Servant Leadership. Servant Leadership means putting the needs of others before your own needs. Instead of asking, “What’s in it for me?” A Servant Leader would ask, “How may I serve you?” At DVT, we put the needs of our customers and partners ahead of our own needs. Agape – (ä-‘gä-pā) Greek Term meaning – Unconditional Love. DVT ‘s Agape Support is Unconditional. Our Vision