IDHA : Images Distribues Htrognes pour lAstronomie Distribution of Heterogeneous Image data in Astro

1
IDHA Images Distribuées Hétérogènes pour
lAstronomieDistribution of Heterogeneous Image
data in Astronomy

• CDS, F.Bonnarel, A.Oberto et all.
• LSIIT, M.Louys, C.Collet
• IAP, P.Didelon, Y.Mellier, E.Bertin
• QUB, F.Murtagh
• ACI-GRID project

2
Images in astronomy

• Metadata definition for image sets in diverse
  archives.
• Multicolor image analysis for object recognition
  and cross identification.

3
Metadata for astronomical images

• Goal
• To provide a model for astronomical image data
• General in content supports the information
  available in different archive servers
  (observatories, surveys).
• Logical in design includes the different steps
  of data production, from raw measurements to
  final products.
• Science-oriented well-suited to user query
  types.
• Strategy
• UML model, XML data description.

4
From Data Model to XML description

• The Data Model
• It is a road-map to identify and describe the
  different kinds of information available on
  images and their relationships.
• Metadata can deal with
• Observation process, instruments,reduction
  process, image storage and distribution.
• http//alinda.u-strasbg.fr/IDHA/lastmodel

5

• The XML documents
• Describe partial views of the DM
• Correspond to user queries
• Object ? resource in a V0table
• Attributes ? Field of a table
• Simple queries ObservingProgram.xml
• ObservingGroup.xml
• ProcObservation.xml

6
Structuring image description

• Collections of images
• ? factorize common information.
• Goods demo goods_idha2.xml

7
IDHA data model development plan

• Sketch out a model description of astronomical
  image data .
• Implement and test the model and XML description
  in the context of the Aladin server.
• Compare/Extend to other archives
• Test it for the AVO/Goods demo(Jan03).
• Ask for more input and criticism.

8
Problems

• Factorize general information for data
  collections (uniqueness, simplicity)
• Allow different levels of detail
• Coarse to fine according to the users goal
  visualise, process, cross-identify, etc
• Data models in the VO context
• Homogenize object names and attributes
• Define a dictionary/thesaurus?
• Tag content with UCDs ?

9
Build in VO evolution

• Data (image) interpretation is the result of a
  process.

• Support versionning, on the fly calibration, ...

10
Conclusion

• Common modeling effort exists in the IVOA. (DM
  Boston meeting)
• Develop compatible metadata objects in the
  UML/XML framework.
• Define the properties of a VO recommended
  datamodel.
• Next steps
• Compare data models and identify common concepts.
• Define some standardized vocabulary and/or
  tagging method to describe common metadata.