Distributed Catalogue Search of Earth Observation Data.

摘要

Catalogues in different organizations use different metadata information models, catalogue interfaces, and transport protocols to support discovery, access and use of Earth Observation (EO) data. The heterogeneity in these catalogues makes it difficult for scientists to find required data. There have been various efforts on integrating distributed catalogues to facilitate EO data search, however, there still lacks an integrated framework that supports EO data discovery on the dataset level and the data granule level simultaneously. This dissertation addresses this gap and designs an integrated catalogue framework (ICF) to integrate distributed catalogues for EO data products. The ICF presented in this study will enhance EO data search by unifying different models and interfaces used in distributed catalogues and releasing users from the heterogeneity of these geospatial catalogues services. This framework not only offers a harmonized interface for users to access distributed catalogues but also supports two levels of search granularities: 1) keywords-based dataset search and 2) data granule search.;To achieve interoperability for the proposed ICF, two levels of mapping are required for each search granularity: One is the mapping of query language to hide users from the various heterogeneous query interfaces; the other is the metadata information model mapping to offer users not only a consistent mechanism to specify search criteria but also unified and integrated search results. This research adopted the OpenGIS Catalogue Service for Web (CSW) specification as the core-underlying standard and leveraged the concept of mediator and wrappers for resources integration. Search requests are translated to CSW Query Language (CQL) and dispatched to distributed catalogues by mediator. Search results from these catalogues are mapped by wrappers to a format compatible with the CSW 2.0.2 - ISO Metadata Application Profile (OGC 07-045) and the Core Profile derived from the Dublin Core Metadata Element Set. A detailed use case for integrating two different catalogues is described to show the capabilities of this framework to support distributed Earth Observation data search on two levels.