Multi-media In Manufacturing

1
Multi-media In Manufacturing

• 2002. 6. 28
• MAI LAB
• Ryu Mi Wyun

2
Manufacturing Today
Current situation at manufacturing companies
What are your biggest problems with global
manufacturing?
Poor visibility into plant
38
operations
Inaccurate demand
36
forecasting
Poor communication
24
Supply shortfalls
18
Poor customer satisfaction
8
Don

t know
8
(multiple responses accepted)
Percent of 50 global manufacturing companies
responding
Forrester Research, July 2000
3
Traditional Supply Chain
Suppliers
Traditional Mfg.
Customers
Loose Coupling
Loose Coupling

• Monolithic Deployment
• Build-to-Stock
• Run-to-Break down
• Replenishment-to-Store

4
e-Manufacturing Transparent Data Access
Emerging e-Commerce Supply Chain
Suppliers
e-Manufacturing
Customers
Seamless Coupling
Seamless Coupling

• Rapid Deployment
• Build-to-Order
• Non-Stop Operation
• Supply Chain Integration

5
What does the e mean?

• The Right Information to the Right People at the
  Right Time
• Keeping the plant in synch with evolving
  business strategy
• Adaptable and Dynamic
• Collaboration with Suppliers and Customers
• Speed and More Speed

e-Manufacturing
6
What does the e mean?

• The Right Information to the Right People at the
  Right Time
• Keeping the plant in synch with evolving
  business strategy
• Adaptable and Dynamic
• Collaboration with Suppliers and Customers
• Speed and More Speed

e-Manufacturing
7
A review of multimedia technology in manufacturing

• A. Gunasekaran , P.E.D. Love
• Department of Manufacturing and Engineering
  Systems, Brunel University
• School of Architecture and Building, Deakin
  University
• Computers in Industry 38 (1999) 65 - 76

8
Contents

• Introduction
• Multimedia and Manufacturing
• Application of multimedia in Manufacturing
• Design of multimedia systems in manufacturing
• Future research
• Conclusion

9
Introduction

• Multimedia melding of text, sound, photos,
  voice and video
• Allows end user to share, communicate and
  process a variety of
• forms of information.

Multimedia Systems
10
Multimedia Manufacturing
Future Information-Oriented, Knowledge-driven
11
Application of Multimedia
12
Design of Multimedia systems

• Improve the integration between functional areas
• Improve the interaction among people
• Reduce the lead time in exchanging information

Central Server (Information Provider)
WAN
Metropolitan Server (Storage Provider)
MAN
Local Server (Access Provider)
LAN
Functional Area
13
Future research Conclusions

• Integration with Internet Telephony.
• Multimedia to minimize waste (scrap, rework,
  etc.) .
• Interactive company websites
• Integration of computer s/w technologies such
  as,
• (expert systems, artificial neural networks,
  hypertext, etc) .

the information flow (Communication)
Multimedia can improve..
the integration of various sub-systems

14
Visual processing and classification of items on
a moving conveyor a selective perception
approach

• H. Isil Bozma , Hulya Yalçin
• Intelligent Systems Laboratory, Department of
  Electrical Electronic Engineering, Bogaziçi Univ,
  Turkey
• Robotics and Computer Integrated Manufacturing 18
  (2002) 125 - 133

15
Contents

• Introduction
• Selective Visual Processing
• Using attentional sequences
• Application Results
• Defective Item Identification
• Automated Sorting
• Conclusion

16
Introduction (1/2)

• Visual Processing and Classification
• Looking at the item on the conveyor via some
  type of sensor(camera)
• Localizing any single item
• Classifying the item based on a set of features
• Performing the necessary action

17
Introduction (2/2)

• The goal
• - To determine the shape of an item under view
  and whether there are any deviations from its
  golden models.

• Issues
• The items to be inspected may contain holes and
  extrusions
• Items shapes may not be regular
• Items positions and orientations may be
  arbitrary
• Real-time visual processing
• Minimal special hardware requirements and
• New items may be added frequently

18
Selective Visual Processing (1/3)
19
Selective Visual Processing (2/3)

• To determine where to look next in the image

1. Initialization
4. Candidate next foveas determine
from
2. Finding a first fovea randomly fixate,
, ,
5. Saliency measure for ,
3. Current periphery
6. Next fovea ,
,
20
Selective Visual Processing (3/3)

• To determine the state of each fovea

1. Initialization get , ,
In 2D shape, ( ,
i indicates an edge oriented )
2. Fovea state ,
3. Adding to the attentional sequence
21
Using Attentional Sequences (1/2)
Contour Identification

• Reference vector using the center of the
  item/radial coordinates

22
Using Attentional Sequences (2/2)

• Identification whether the item matches a
  model from the memory is made

1. Euclidean norm of the difference compute
4. Compute average position error
2. Comparing the difference
try next model, the item
is type
5. Using the measure if ,
Perfect match Otherwise, missing subparts
3. Identifying corresponding subparts
23
Application Results Defective item
identification

• For a subcontracting firm manufacturing metal
  door parts
• To determine whether any has missing holes or
  not.

24
Application Results Automated Sorting

• Automated remote controller sorting system

25
Conclusion

• Approach to visual processing based on selective
  perception
• Defective Item Detection
• Automated Sorting
• Flexible and Real-time