Jérôme Gasperi
CNES activities on semantic search
CEOS WGISS-35
INPE - Sao Paulo dos Campos, Brazil - May 7th, 2013
Slide 2
Slide 2 text
Semantic search
Linked Data
Faceted search
Slide 3
Slide 3 text
Semantic search
Help users to find the right data
Slide 4
Slide 4 text
AMMA
CYCLOPES
GEOLAND
GLOBCOVER
PARASOL
POLDER
....
Slide 5
Slide 5 text
AMMA
CYCLOPES
GEOLAND
GLOBCOVER
PARASOL
POLDER
....
Land cover
Soil moisture
Temperature
Reflectance
Vegetation
...
Slide 6
Slide 6 text
How ?
Slide 7
Slide 7 text
Tag metadata at the product level with keywords
related to the project or mission name
e.g. search for "Land cover" products instead of "GLOBCOVER"
Slide 8
Slide 8 text
Use Exogeneous data to characterize product content.
The product footprint should be use for automatic processing - e.g. toponyms extraction with OpenStreetMap
Slide 9
Slide 9 text
Example
Using Global Land Cover 2000 to automatically characterize product thematic content
Slide 10
Slide 10 text
Global Land Cover (GLC 2000)
Slide 11
Slide 11 text
Thematic sliders
SPOT cover - 1,300,000 scenes
Slide 12
Slide 12 text
Search for coastal cultivated area ?
Slide 13
Slide 13 text
Search results
(footprints displayed in black over the map)
Coastal cultivated area
Slide 14
Slide 14 text
Thematic classification
is displayed within
the result metadata
costal...
...cultivated area
Slide 15
Slide 15 text
https://vimeo.com/51045597
!
Slide 16
Slide 16 text
Linked data
Achieve multidisciplinary interoperability
Slide 17
Slide 17 text
« Linked data is the right way to do Semantic Web »
Tim Berners-Lee
Slide 18
Slide 18 text
The Graph of Idea - www.griffsgraphs.com
Slide 19
Slide 19 text
Faceted search
Break the hierarchy of search filters
Slide 20
Slide 20 text
Web search paradigms
Taxonomy search
(e.g. Yahoo! Directory)
Direct search
(e.g. google)
vs.
Slide 21
Slide 21 text
Faceted search
Taxonomy search
(e.g. Yahoo! Directory)
Direct search
(e.g. google)
+
(a sort of)
Slide 22
Slide 22 text
« Faceted search enables users to navigate a multi-
dimensional information space by combining text search
with a progressive narrowing of choices in each dimension »
http://en.wikipedia.org/wiki/Faceted_search