Databases Reference
In-Depth Information
geometric shape such as points, lines and polygons. Points are represented as
pairs of latitude and longitude coordinates, lines as strings of coordinate pairs,
and polygons as lines that form closed loops or areas. In addition, the vector
model can represent the topological attributes between two geometries, such
as adjacency and containment. For instance, a water body or a zoning area
can be represented with a polygon. Thus the spatial attributes record data
about the location, topology and geometry of geospatial data.
The second component in the vector data is the non-spatial attributes, also
called thematic attributes, that refer to non-spatial properties of geospatial
data, such as annual rainfall, vegetation type, zoning type, land use, states,
census tracts, etc. A thematic layer is a collection of geometries having the
same attribute set. Thus we can have a layer of schools as points, and another
layer describing roads and bridges as lines and points.
2.2 Raster data
Under the raster data model, the spatial data, such as satellite images, eleva-
tion maps, or digitized maps, is represented as a grid of columns and rows, i.e.
as a matrix of cells (called pixels). Each layer of grid cells in a raster model
records a separate attribute. Each cell carries the non-spatial data, such as
rainfall, temperature, vegetation type, etc. Spatial coordinates are not usually
explicitly stored for each cell, but implicitly represented with the ordering of
the pixels. Typically, each layer contains information about the number of
rows and columns and the geographic location of the origin. The spatial reso-
lution of a raster is the size of one the the pixels on the ground. For example, if
one pixel corresponds to 3 meter by 3 meter area on the Earth, the data has 3
meter resolution. Different sensors provide images of different resolution levels
[24]. For instance, the Advanced Very High Resolution Radiometer (AVHRR)
produces images of 1km resolution [20], the Landsat Thematic Mapper (TM)
multi-spectral images of 30 meter resolution, the Radar images of 5 to 10 me-
ter resolution, and the IKONOS satellite images of 1 and 4 meter resolution.
DigitalGlobe's QuickBird satellite's panchromatic images of 60 cm resolution
marks the world's highest resolution commercial satellite imagery.
Satellite images also carry temporal information, designating either the
time when the image data was downloaded from a satellite, or the time of
data creation. The coverage area of images are different from image to image,
depending on the orbits of the polar-orbiting satellites. The satellite images
also require georegistration that registers each image with a known coordi-
nate system (e.g. longitude, latitude) and reference units (e.g. degrees), and
assigns coordinates to the left, right, top and bottom corners of the image.
One satellite may have several sensors capturing information in different fre-
quency bands. The information in these bands can be further processed bands
to produce different layers, such as water, vegetation and the like.
