KIM Knowledge Information Mining

1
KIM Knowledge Information Mining

• Andrea Colapicchioni
• Advanced Computer Systems Space Division
• a.colapicchioni_at_acsys.it

2
Image understanding

• In general, by image we understand picture thus
  relating it to the (human) visual perception and
  understanding
  
• A picture is characterized by its primitive
  features at different scales, such as
  
• Colour
• Texture
• Shape

3
EO Images are signals

• The concept of EO image is beyond the pictorial
  understandings
  
• EO Images are multidimensional signals obtained
  from sensor like
  
• Radar
• Optical Multi-spectral
• Optical Hyper-spectral

4
KIM Knowledge Driven Information Mining
Shape
Clustering
Primitive Feature Extraction
Images
Texture
Clustering
Spectral
Clustering
5
From data to semantics
6
Image Co-registration Experiment
2000 x 2000 Landsat
A Landsat scene is co-registered on ERS
2000 x 2000 ERS
ERS Scene
Co-registration
Landsat Scene
2000 x 2000 Landsat
2000 x 2000 ERS
Generating from 15 to 20 ErsLandsat couples
using homologous point and warp (with ENVI tool)
the images are cut in 2000x2000 tiles
7
KES Catalogue knowledge