Temel donanım bilgileri
© Copyright Brian Brown, 1992-2001. All rights reserved.

INPUT/OUTPUT DEVICES CONTINUED

Optical Readers (character, page, bar code, MICR etc)

• convert printed or hand written information to computer data
• information is read in one three ways
  • OMR
    marks which are placed in predefined areas of the document
  • OCR
    printed or typewritten characters
  • OCR
    handprinted characters
• pre-printed paper is normally used for hand-printed marks or characters
• an example of OMR is LOTTO
• readers are designed to read marks or characters positioned in a matrix pattern
• colored printing is used in pre-printing, which is insensitive to the reader

MARK READERS

• detect the presence or absence of a mark in a specific position

  
  Fig 6.11: Mark Reader

• simpler optical mark readers require guidelines

  
  Fig 6.12: Mark Reader showing guidelines

• generally, marks can be made using pencil, ballpoint pen etc (some colors might not be recognized, like red)

PRINTED TYPE/CHARACTER READERS

• OCR-A
  • machine readable fonts are used

    
    Fig 6.13: OCR-A font
    
    

  • first to be produced, originated in USA
  • characters designed for machine recognition
    
    
• OCR-B

  
  Fig 6.14: OCR-B font
  

  • originated in Europe
  • less stylized, more characters than OCR-A
  • by reducing the number of characters in the character set, recognition is made quicker, easier and more reliable
    
    
• hand printed characters
  • differences in individual writings make this difficult
  • many OCR machines restrict to numeric digits
  • often shows desired input on pre-printed form
• Performance
  • typical speeds 300cps-800cps
  • error rate depends on quality of input and size of character set used
  • typical reject rates on hand printed characters of 1%
  • software often uses intelligent substitution
• Applications
  • alternative to keying in data
  • opinion polls, market surveys
  • tests, stock control, lotto

Optical Character Recognition [OCR]
OCR is the scanning of text documents into graphic images, then using software to decode the graphic picture elements back into text.

1. When the scanner scans the document, it is read as a series of black and white pixel (dots) elements. This process often tends to degrade the edges of the text characters, and is more pronounced when the characters on the original are too small. Edge degradations makes it harder for the OCR software to convert the pixel elements back into text later on.
2. The OCR software reads the bitmap of pixels created in step 1 and averages out the white spaces on the page, effectively identifying paragraphs and eliminating graphics. The white spaces between each line of text is used as a baseline reference for recognizing the characters on that line.
3. First, the OCR software tries to match each character on a line in the bitmap against character templates that it knows about.
4. The remaining unidentified characters have a technique known as feature extraction applied to them. The OCR software calculates the characters height, lines, curves and other features. It can then make close guesses as to the characters value.
5. For the remaining characters the OCR software cannot recognize, the software can either apply contextual analysis, which basically means looking at the syntax and construction of the words and making a guess (for example, changing thi5 to this), or give up and substitute the unknown character with a distinctive symbol such as ~ or @.
6. The finished information is normally able to be saved in a number of different formats, text or Rich Text Format (RTF). OCR software which support RTF can also recognize bold, italics, retain tabs and whitespace, as well as recognize a limited number of different fonts.

Bar Codes

• represents numeric data as a series of bars
• bars have varying thickness and separations
• numeric data is often written underneath the bar-code
• easily read by light-pen or scanner device
  The light pen has a sensitive tip which contains a light source and light detector. When the pen is stroked along the bar-code symbol, the light from the pen bounces off the dark bars to produce a corresponding set of binary pulses. This sequence is decoded to give the numeric data that the bar-code represents.

  
  Fig 6.15: Bar-Code symbol

• Examples of Bar-Code Applications
  
  • consumer goods (super-markets)
  • stock inventory
  • library systems for cataloging books
  • drug dispensing at your local pharmacy
    
    
• Bar-Code Details
  
  • each digit is made up of two black and two white bars in alternate sequence
  • the widths of the black and white bars add up to a fixed width for all characters
  • each digit character can be broken down into seven elements

    
    Fig 6.16: Bar-Code encoding details

  • a digit is coded differently depending upon whether it is in the right half or left half of the bar-code symbol
  • all digits on the left have have an odd number of black bars and begin with a white bar
  • all digits on the right half have an even number of black bars and begin with a black bar
  • this permits scanning in either direction
  • the two halves of the bar-code symbol are separated by two black guard bars
  • two black guard bars are used to mark the beginning and end of the bar-code symbol
  • Universal Product Code (UPC) contains a manufacturer code and a product number
    
    
• Performance
  
  • error rates are very low
  • bar codes are easily and quickly re-scanned
  • blemishes in the bar-code causes errors
  • anything that reflects light causes problems, e.g., ice.
    
    
• Examples of Bar-Codes
  

  
  
  Fig 6.17: Bar-Code Examples
  

Typical Connections for a Bar-Code Reader
The following diagram shows the bar-code reader inserted between the keyboard and the base unit of the computer. The bar-code unit converts the information on the bar code information read by the pen presents it to the computer as a series of ASCII characters. The computer thinks that the characters came from the keyboard. This simplifies writing software to read bar-codes.

Fig 6.18: Bar-Code Connections

Typical Bar-Code Reader for a PC
The following diagram shows a typical bar code reader unit for a Personal Computer.

Fig 6.19: Bar-Code reader and pen

Summary
Optical readers consist of OMR and OCR. An example of OMR is LOTTO.

Optical Character Recognition is achieved by scanning the image is as graphics, and then uses a series of comparisons to try to convert each detected character area to a known character. Problems occur when the font is too small, it is skewed at an angle, or insufficient contrast exists between the background and the characters.

Bar codes are in extensive use today. They are primarily used for identifying products. They are easily scanned and stored in a computer.

Home | Other Courses | Notes | Tests | Videos
© Copyright Brian Brown, 1992-2001. All rights reserved.