Browse Prior Art Database

METHODOLOGIES AND IMAGE PROCESSING TECHNIQUES FOR AUTOMATIC MARK/LABEL VERIFICATION AND RECOGNITION

IP.com Disclosure Number: IPCOM000006045D
Original Publication Date: 1991-Apr-01
Included in the Prior Art Database: 2001-Nov-28
Document File: 1 page(s) / 55K

Publishing Venue

Motorola

Related People

Christopher J. Le Beau: AUTHOR

Abstract

Graylevel digital image processing techniques are employed in automatic visual interpretation and recognition of product marking and semiconductor package inspection. A mark or label is a group or region of n alpha-numeric characters or patterns, including logos and other symbols. Any solid- connected contiguous blob is defined as a single character within the label. The characters can be darker or lighter than the background on which they appear.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 64% of the total text.

Page 1 of 1

0 M

MOTOROLA INC. Technical Developments Volume 12 April 1991

METHODOLOGIES AND IMAGE PROCESSING T@HNlQUES FOR

AUTOMATIC MARK/LABEL VERIFICATION AND RECOGNITION

by Christopher J. Le Beau

  Graylevel digital image processing techniques are employed in automatic visual interpretation and recognition of product marking and semiconductor package inspection. A mark or label is a group or region of n alpha-numeric characters or patterns, including logos and other symbols. Any solid- connected contiguous blob is defined as a single character within the label. The characters can be darker or lighter than the background on which they appear.

  Verification and recognition of a mark/label is achieved by an ability to automatically judge the readability of the entire label and individual character. Readability is tested by a non-font based approach by transforming the label through such operations as scaling, rotation, and individual character connected segment variations. The label location and transformation is identified by location of hvo or three spatially unique interior label characters, called reference characters. The spatial relationship among these reference characters determines the label's

transformation, and thus the location at which the individual label characters must reside.

  Character readabilityiis determined using direction edge correlation techniques using comparisons to taught character edge-shapes. Direction edge information is obtained by traditional gra...