Section: New Results
Ontological description of a remote sensing processing chain
The Clime team is studying the use of ontologies for documenting complex remote sensing processing chains.
A first objective is to provide end-users with a documentation adapted to their level of expertise and corresponding requirements. A first group of users includes scientists in remote sensing or environmental experts or geologist, etc . They require a step by step execution of the processing chain, in order to analyse and check intermediate results, change parameterisation, etc . On another hand, operational users would prefer a black box with fully automatic execution using sensible default values, etc .
A second objective is to take advantage of the applicability and usability of ontologies: ontologies help for adapting a given processing chain to another site and/or another application; they enable to share knowledge between different users/developers of the processing chain: a change in the processing chain is documented and described by the ontology; they provide means to reason about knowledge, e.g. at the highest level on the processing chain, at the lowest levels, on data and site specific issues.
These objectives have been fulfilled by developing a hiearchical ontology, based on a tree structure, for describing a given processing chain: land degradation monitoring from MODIS data, developed in the context of the ENVIAIR project. We hope that this ontology will help adapting this processing chain to other applications, such as desertification monitoring (DESMED P3+3 project).
Current research is geared at the automatic derivation of ontologies, consistent with the available software components.