Image Information Mining Services

1
Image Information Mining Services

• SSE 2004 Workshop
• Service Support Environment an introduction
• ESRIN, Frascati April 22, 2004
• Andrea Colapicchioni - Advanced Computer Systems
  Rome

2
IIM Roadmap
3
The systems we made

• KIM - Knowledge driven Information Mining in
  remote sensing image archives 2001-2002 DLR, ETH
  Zurich, ACS Rome, NERSC Bergen
• KES - EO domain specific Knowledge Enabled
  Services 2002-2004 ACS Rome, DLR
• KIMV - KIM Validation for EO archived data
  exploitation support 2003-2004 ACS Rome, DLR

4
KIM System

• KIM implements a brand new concept to explore
  image catalogues
• It permits to search in an archive by giving
  examples directly on the images, implementing
• CBIR (Content Based Information Retrieval)
• Interactive IIM (Image Information Mining)

5
Architecture
Primitive Features - Texture - Colour - Shape
Interactive Learning (Using Bayesian Networks)
Features
Images
Primitive Feature Extraction
River
010010 101101 110010
010010 101101 110010
Forest
Cloud
DATA
INFORMATION
6
Interface
7
KIM System application

• KIM implements a brand new concept to explore
  image catalogues
• It permits to search in an archive by giving
  examples directly on the images, implementing
• CBIR (Content Based Information Retrieval)
• Interactive IIM (Image Information Mining)

8
An application of Kim(cloud cover
characterization)
MERIS Level 1 Reduced Resolution
cloud
KIMV System
9
An application of Kim(cloud cover
characterization)
Map of Cloud object
10
An application of Kim(cloud cover
characterization)
Binarised, closed MAP of cloud feature
11
An application of Kim(cloud cover
characterization)
12
Integration in MASS

• Through the toolbox the KIM database has been
  easily linked as Simple Catalogue Service to the
  MASS portal via a JDBC connection
• KIM advanced information extraction functions are
  immediately available in a wide environment

13
MASS
14
System Context
15
Future opportunities

• Identify images on the basis of their content
  (fires, algae bloom, urban)
• Features (according to OGC standards)
• Thematic Maps

16
Requirements - 1

• In the Simple Catalogue Service not all search
  conditions can be expressed, thus limiting the
  search
• The collection concept has impact on easiness of
  use and performance
• It is necessary to link search results to order
  handling

17
Requirements - 2

• Display area of interest on thumbnails and
  quicklooks
• The system should be able to show features on the
  map
• Search multiple features should be possible

18
Requirements - 3

• It should be possible to search features also via
  spatio-temporal proximity queries (example ship
  and oil spill)
• Multi-use of services should be addressed

19
Requirements - 4

• Access control and e-commerce issues should be
  identified and solved
• Intellectual property right issues might emerge
  in future

20
Contacts

• Andrea Colapicchioni
• KES, KIMV project manager a.colapicchioni_at_acsys.
  it
• Prof. Mihai Datchu
• DLR Oberpfaffenhofen
• mihai.datchu_at_dlr.de
• Prof. Klaus Siegel
• ETH Zurich
• siegel_at_vision.ee.ethz.ch