The Human Geography Knowledge Discovery (HGKD) is challenged by the sheer scale of “information
items” (documents, web pages, images, voice, video, mapping, imagery, etc.) that either exist now
or are projected in the near future. In particular, the quantity of geospatial/image/video data is
growing very rapidly.
Semantic technologies (e.g., using OWL/RDF) provide an elegant representation for HG
objects/concepts, relationships, and events. These provide a means of describing image content both
in terms of pure image-organization and interpreted objects. These methods can be extended to
representing activities within video segments. However, use of this more expressive formalism
requires substantial additional computation. Given the existing and projected data volumes, there
is a computational challenge in providing rich and detailed representations to all data, and to
using the full richness of these representations during KD.
A comprehensive KD solution based on an Enterprise Data Layer can be developed by adopting a
multi-layered, combined top-down / bottom-up representation and control system, incorporating
multiple existing representation solutions. Benefits of this approach include:
Available tools, processes, and services for use in a human geography workflow:
The HG Enterprise Data Layer
The HG enterprise Data Layer will provide control of mission critical information to the data owner
and make information manageable, portable, and easier to share. Mission critical information,
automatic provisioning of analyst tools, and secure information sharing through existing tools will
enable community of interests (COIs) portals, information sharing services, legacy integration,
federated metadata catalogs, databases, and enterprise services.