32
Track, Trace & Control Solutions © 2011 Microscan Systems, Inc. Learn about OCR: Optical Character Recognition

Learn about OCR: Optical Character Recognition

  • Upload
    dawson

  • View
    66

  • Download
    1

Embed Size (px)

DESCRIPTION

Learn about OCR: Optical Character Recognition. About Your Presenter. Presenting today: Juan Worle Technical Training Coordinator Microscan Corporate Headquarters Renton, WA. Course Objectives. By completing this webinar you will: Understand definition of OCR - PowerPoint PPT Presentation

Citation preview

Page 1: Learn about OCR:  Optical Character Recognition

Track, Trace & Control Solutions© 2011 Microscan Systems, Inc.

Learn about OCR: Optical Character Recognition

Page 2: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

About Your Presenter

Presenting today:

Juan WorleTechnical Training CoordinatorMicroscan Corporate Headquarters

Renton, WA

Page 3: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

Course Objectives

By completing this webinar you will: Understand definition of OCR

A little about the history, and where it is applied today

Know different types of OCR and how OCV is different Understand how to select the best tools

Know the critical features of OCR fonts Learn how to identify potential weak points in an application

Know how to identify reliable OCR applications Become familiar with applications that have been successful and low maintenance

Page 4: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

Topics

About OCR OCR and OCV Decoding OCR Example applications

Page 5: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

About OCR

What does OCR mean, and some perspective

What is OCR? A little history OCR and Machine Vision

Page 6: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

What is OCR?

Optical Character RecognitionThe conversion of written or typed text into a string of characters formatted

for machines.

Page 7: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

What is OCR?

Optical Character Recognition

OCR fonts are unique: Unlike barcodes and 2D symbologies, they are both machine readable and human readable.– The data is considered less secure than barcodes and 2D symbols.

Some OCR software tools convert paper documents to electronic documents.

OCR conversion on a PC allows you to copy scanned text

Page 8: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

A Little History

OCR has been used commercially since the 1970s. Automated bill processing: OCR systems in automated payment

processing facilities Retail check-out before UPC: Handheld OCR readers read the

price of merchandise

Automatic check processing machines use OCR algorithms and MICR fonts

The first patents were developed in the 1930s by Gustav Tauschek and then Paul Handel

Page 9: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

A Little History

Today OCR is used in many specialized applications. Search engines Handwriting recognition Postal tracking and document handling

Google’s powerful OCR software allows you to search the web

from a mobile phone

Mailing systems use specialized OCR algorithms for handwriting recognition

Page 10: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

A Little History

OCR within Machine Vision focuses on industrial applications. Automotive, aerospace, semiconductor manufacturing Food and beverage handling Packaging

Industrial applications for OCR have the following traits:Fixtured partsConsistent lighting and environmentConsistent fonts

Example of LOT and DATE codes:By reading the text, the date can be checked, and the lot number verified

Page 11: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

OCR and Machine VisionThere are three uses for OCR and OCV tools.

OCR is used to identify the contents of unlabeled cans

Presence: Ensure the characters have been marked– Ensure the characters are present– Check the readability of OCR characters– Optical Character Verification (OCV) is common

Tracking: From stock through manufacturing to packaging– Lot, batch, expiration dates, serial numbers– A common barcode application

Identification: Identify part or contents of a container– Ensure proper labeling– Ensure product matches container

Page 12: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

OCR and OCV

Understand the difference between Recognition and Verification

Comparison of OCR and OCV Methods to read OCR

Page 13: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

Comparison of OCR and OCV

OCV: Optical Character VerificationUse OCV tools to check the legibility and quality of text, based on a fixed and

known sequence of characters.

The output of an OCV tool is a quality report of correctness.

OCR: Optical Character RecognitionThe OCR tool is used to read an unknown sequence of characters.

The output of an OCR tool is machine usable text.

OCV: verify quality

OCR: read text

Page 14: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

Comparison of OCR and OCV

Verification:Inspecting characters for content, correctness, quality, contrast and sharpness

compared to stored templates.

Examples: Date / Lot verification Component ID verification

– Label, carton, insert, outsert Verification of on-line printing

– Clinical labels– Blister packs – Direct printing on product

ABC-123

Use OCV to check levels of print quality and legibility

Page 15: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

Comparison of OCR and OCV

Reading: A tool for reading text strings of random content and converting to machine

usable text.

Examples: Sorting and identification Serialization Codes with Time/Date stamp Verification of readability

– decoded=readable Verification of on-line printing

– Codes with inconsistent character placement or size– Can verify text using match

OCR can be used to read serial numbers on a data plate

Page 16: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

Methods to Read OCR

Fixed Font:The font characters must conform to a fixed pattern.

Examples:– OCR-A, OCR-B: Many printed applications such as passports, documents, and

pharmaceutical labels– SEMI: Used for semiconductor manufacturing– MICR: Banking documents such as checks.

Page 17: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

Methods to Read OCR

Trainable Font:Any font can be presented and learned by Machine Vision software during

setup, then identified during run-time.

– More common than fixed font because any font or variations can be trained.– Reviews each character and looks for a match in the trained font library.

When using trainable font tools, a character is not recognized until it is trained

Trainable OCR tools let you use a non-standard font

Page 18: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

Decoding OCR

Understanding the unique traits of an OCR font

Character dimensions Font characteristics Print variations Improve performance

Page 19: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

Character Dimensions

The overall size of the character matters, as well as the features.

AHeight

Width

Features

Area

Line Weight

Page 20: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

The font matters. Maximize the difference in similar characters for more reliability. Many characters have very little difference.

OC Q G

Font Characteristics

OLet’s look at the Arial font. The Character O has many similar characters:

O OOriginal Letter C

80% matchLetter Q

75% matchLetter G

70% matchArial: High probability of confusion

Page 21: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

Fonts designed for machine reading work best. Uniform character spacing Each character designed to be different than all others DPM applications

Font Characteristics

Low probability of confusion: OCR-A is designed for machine reading and has differences in similar characters

Verdana Sample

OCR Sample

Even character separation improves readability

There are also several fonts designed for Inkjet and Direct Part Mark (DPM) applications

Good:

But better:

Page 22: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

Print Variations

Even if the text looks good on screen, printers can change appearance.

SKEWANG EL

SCALE

LINE WEIGHT

DEFECTS

Print considerations DPM considerations

Dot Size

Overprint Underprint

Dot Spacing

Skew

Dot offset

Tip: Avoid large gaps when marking characters.

Page 23: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

Print Variations

The substrate (the material you are printing on) can affect readability.

Ink absorption Background noise Damaged characters

Background noise can cause character confusion

Damaged characters and uneven surfaces can affect decodability

Page 24: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

Improve PerformanceThere are many ways to improve performance. Use trainable font tools for more tolerance

– Teach variations of a font– Line weight

Leave a quiet zone that is 2-3x character space Use additional Machine Vision tools

– Morphology: modify the image– Dynamic location: use an anchor point

Morphology: Vision tools that can improve the appearance of an image

Teach variations of the font for more reliable reading

Dynamic location is helpful if the part location moves

De-focus

Line weight

Line weight

Original

– Slight rotations– Focus

Page 25: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

Improve PerformanceRecommendations when using Microscan products. Use at least a 6 point font size Adjust camera to font for 25 pixels wide/30 pixels high Space between each character should be at least 1 point (0.015”)

– Half the size of the character works best The smallest features within a character (like Line weight) should be at least

1 point (0.015”)

A BIdeal space between characters

is half the character size

Feature size > .015”

25 Pixels

30 Pixels

Page 26: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

OCR Applications

Some common OCR applications

Print verification Label verification Date and Lot code tracking Part identification

Page 27: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

Print Verification

Continuous ink-jet (CIJ) on cartons. Validate the data Combine with barcode tool Feedback when the head should be cleaned

Page 28: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

Label Verification

Ensure the proper label is applied. Several products are run on a single line Report error when incorrect label is applied

Page 29: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

Date and Lot Code Tracking

Date and Lot code traceability. Validate and verify printing Track products through manufacturing Conform to regulations (FDA)

Page 30: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

Part Identification

Identify gasket for installation.Read the part number to ensure the correct part is installed

Page 31: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc. © 2011 Microscan Systems, Inc.

Learn about OCR

Conclusion The idea of OCR is not new; it has been around since the 1930s. Industrial

applications gained momentum in the 1970s. Today OCR tools are used in many non-industrial applications. Machine Vision

OCR tools focus on industrial applications. OCR tools can be categorized by OCV, OCR Fixed Font or OCR Trainable Font. The features of a font are important and can determine the success of an

application.– Font selection– Probability of confusion– Printer variables

The most common Machine Vision applications include Presence, Tracking, and Identification.

– Substrate and marking method– Character separation– Character dimensions

Page 32: Learn about OCR:  Optical Character Recognition

Learn about OCR

© 2011 Microscan Systems, Inc.

Thank you!For more informationWebsite: www.microscan.com

– Online courses– Spec sheets– Technology brochures– Support self-help and support request form

Instructor:Juan Worle, Technical Training CoordinatorEmail: [email protected]

Feedback on this webinar: www.microscan.com/feedback

Additional contacts:Product information: [email protected]: [email protected]: [email protected]

Graduation exerciseDownload Visionscape from www.microscan.com download center