Browse Prior Art Database

ROTATIONAL, ADAPTIVE, INTERPRETIVE, MACHINE VISION ALGORITHM

IP.com Disclosure Number: IPCOM000006490D
Original Publication Date: 1992-Aug-01
Included in the Prior Art Database: 2002-Jan-09
Document File: 2 page(s) / 90K

Publishing Venue

Motorola

Related People

David Karpinia: AUTHOR [+2]

Abstract

Inspection of cosmetic and scratch defects is purely objective at the operator level. The tools available are limited to training, physical examples, modules, and per- sonal experience of the operator. Numerous problems exist with possible defect passage with operator change and correlation between published specifications and those actually in effect due to human perceptions. To quantify the inspection criteria and attempt to regulate the material some form of automated inspection is nec- essary. The following algorithm was developed to meet this need.

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 50% of the total text.

Page 1 of 2

MOTOROLA INC. Technical Developments Volume 16 August 1992

ROTATIONAL, ADAPTIVE, INTERPRETIVE, MACHINE VISION ALGORITHM

by David Karpinia and John Christopher Lewis

  Inspection of cosmetic and scratch defects is purely objective at the operator level. The tools available are limited to training, physical examples, modules, and per- sonal experience of the operator. Numerous problems exist with possible defect passage with operator change

and correlation between published specifications and those actually in effect due to human perceptions. To quantify the inspection criteria and attempt to regulate the material some form of automated inspection is nec- essary. The following algorithm was developed to meet this need.

  The algorithm is three stages; the first part utilizes a fudical with perpendicular intersection of two lines to determine orientation axis of the viewed image refer- ence to that ofthe stored image. The second determines if mechanical rotation is necessary. The third part will selectively allow for software manipulation to simulate rotated conditions of the two images. The choice of mechanical rotation is only an option if the hardware exists in the system, but is allotted for within the soft-

/

ware. When the product delivery tolerances are tight enough, the algorithm will match the fudicals of the two images and subtract corresponding parts of the images utilizing image arithmetic. The image arithmetic routine returns the absolute value of the ditference of two images on a pixel by pixel basis. The two images involved are the stored image and the viewed image. The stored image is an acceptable specimen in the system memory and the viewed image is the one processing is to take place on. Therefore, the image arithmetic routines currently available are only accurate if the two images are exactly the same, by size, shape, and/or orie...