I would argue strongly that the primary geographic reference should be in latitude and longitude, the only truly global reference system. The problem of georeferencing data is a difficult one, and requires quite a high degree of flexibility. In work at British Antarctic Survey, the following types of georefernced objects have been identified:
Point
Line (defined by a series of ordered coordinates)
Box (defined by maximum and minimum latitude and longitude)
Polygon (defined by a series of ordered coordinates)
Point Cloud - a set opoints, each of which is defiend by a coordinate pair.
This has been implemented in the BAS MetaData Management system (http://www.antarctica.ac.uk/mdms/), which may give useful pointers to how this has been done in the past. The underlying technology here is the projection software proj (http://kai.er.usgs.gov) and a Perl module for manipulating GIF images, put together in a series of Perl CGI scripts I didn't do it, so don't ask for details!
Paul Cooper
>>> "Ohanlon, Christopher" <[log in to unmask]> 12 September 2000 10:03:32 >>>
Hi
I am currently working on a project looking at the feasibility of having a
Geo-data Browser for the HE community. At the moment the project is in its
initial stages, and I am trying to find out what level of detail the
academic community would require for a geo-spatial portal.
Firstly let me explain what the portal will do.
Increasing amounts of geo-spatial data are being created within the UK HE
community, much of which has value both within and outside academia. It is
estimated that over 80% of all scientific data sets are geographically
referenced in some way. At the same time the demand for access to
geo-spatial data is growing, but there are two major barriers confronting
the potential use of geo-spatial data resources.
* How to find out what geo-spatial data exist.
* Having located it how to ascertain its quality and suitability for
use.
The solution to overcoming these barriers is to provide comprehensive,
standardised metadata, available through a web-searchable database. The
metadata has to be easily accessible and provided in such a way that an
individual can obtain sufficient information to compare suitability of data
from different sources.
What I am interested in finding out is what level of detail within the
metadata the academic community would require in order to suit its needs.
For instance would the name of the country the dataset is from be enough?,
or would the dataset have to be more specifically referenced, such as the
county, town, administrative area etc? Would it be useful to reference the
dataset by a series of co-ordinates, that could be located on an OS map?, or
is this level of detail too specific? (remember, there could be hundreds of
datasets that would need searching if the geographic reference was too
broad) Also what other information about the dataset would be required to
ascertain what information the dataset contained? These could be Title,
Abstract, Keywords, Author(s) etc
Please could you get back to me with your thoughts on this, and your help
would be greatly appreciated.
Many Thanks
Chris O'Hanlon
Geographic Information Officer
Geo-Spatial Project
The Data Archive
University of Essex
Wivenhoe Park
Colchester
Essex
CO4 3SQ
For more information on the Geo-data Browser project and the associated
Geo-Crosswalk project please go to
http://www.data-archive.ac.uk/news/GeoSpatialData.asp
Chris O'Hanlon
Geographic Information Officer
Geo-Spatial Project
The Data Archive
University of Essex
Wivenhoe Park
Colchester
Essex
CO4 3SQ
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|