|
TerrainWorks© is ideal
for database applications involving storage and manipulation of
geo-spatial and attribute data that describe objects such as roads,
legal boundaries, and streams. It stores groups of point, line, and
polygon data in a compact form that facilitates fast searching.
Management of Terrain-Embedded Linear Networks
The TerrainWorks© DataBlade has been designed to provide spatial data management
professionals with a core tool for modeling, maintaining and querying large, seamless,
linear networks that may be embedded on a terrain surface. Hydrological networks, road
networks and electric power grids are all examples of this type of embedded network.
To achieve truly seamless coverage, all geographic data 3D elements are stored as
(Latitude, Longitude, Height), with computation performed on the ellipsoid. Projection to
a particular mapping plane is done ‘on the fly’, when required. The basic
TerrainWorks© DataBlade data objects, which include versioning and accuracy information
at the feature level, are: - point, line and area features:
Some environment-oriented
examples of these are, respectively, well site, stream segment and forest district. More
urban-oriented examples might be locations of voting sites, streets, and property
boundaries, respectively.
- partitioned coverages: Watersheds, forest-cover polygons,
electoral districts, or municipal boundaries are naturally represented by this type of
data object.
- linear networks: Collections of point, line and area features
can be formed into network configurations. By using new innovative indexing methods, the
TerrainWorks© DataBlade can manage extremely large networks, such as the Fraser River in
British Columbia, and allow fast response to network traversal queries (e.g., upstream and
downstream inquiries).
- datachips: This is an abstract
data type designed by BCS to store data in a very compact form that also supports fast
access. A datachip preserves a user-specified precision on the input data (say 0.5
meters), but stores feature geometry in a compact and geographically partitioned form. The
need for this partitioning arises for features such as Williston Lake in British Columbia,
which stretches across more than 100 of the 7000+ map sheets that cover BC (BC is an area
larger than the States of Washington, Oregon and California combined). To examine a small
area on the shore of Williston Lake in an interactive viewer, it is clearly undesirable to
have to download the entire feature! With datachips, only those datachips that contain the
parts of the feature in the area of interest are downloaded.
Using these core data objects with a set of standard operations (distance, area,
buffering, overlay, intersection, etc.), users can implement and maintain large spatial
data holdings.
The following pseudo-query examples illustrate two key aspects of the TerrainWorks©
DataBlade. These features are: 1) queries directed to interactions between
networks, or between networks and the terrain surface, and 2) version control.
1) Consider the following pseudo-query:
select Area(Watershed(streamElement,
(Intersection(streamElement, roadElement))))
from streamNetwork, roadNetwork
where ContainedIn(Intersection(streamElement,roadElement), userDefinedArea);
This query returns the area of the upstream watershed for each road-stream crossing.
The returned table could then be joined (via position) to maintenance records of road
construction, identifying possible problem spots (e.g., points where a road crosses a
stream with a large watershed).
2) A major element that is missing in the traditional approach to maintaining
geographic data in a GIS system is version management. Road and stream networks are
constantly changing, and managing these changes, including ‘rolling back’ to a
previous epoch, is a key feature of the TerrainWorks© DataBlade. Each
spatial element includes a date of inclusion and a date of retirement. The example below
shows how such information may be used.
select roadElements from roadNetwork
where ContainedIn(roadElement, userDefinedArea)
and dateOfInclusion(roadElement) > lastInventory;
A similar mechanism is used to maintain data with multiple spatial accuracy within a
single database.
Other, more traditional, mapping interfaces are achieved by combining the
TerrainWorks© DataBlade with other IBM Informix GIS DataBlades.
|
Terrainworks
