|
Current procedures for storing 3D geospatial data in spatial databases and GIS systems
can be inefficient, both in terms of their data compression, and in their schemes for
extracting data for a given region. Our newly-developed method for geospatial data storage
and retrieval overcomes these inefficiencies by introducing an intermediate level of
object for data representation - the datachip.
The datachip method allows for very efficient storage of geospatial data by grouping
nearby points, and representing successive points in terms of their integer increments
from previous points, on a virtual grid. Using this scheme, a four-fold decrease in
storage requirements can typically be achieved. The datachip method also provides for fast
extraction of data for a selected region, by pre-generating a multidimensional index for
the bounding boxes of the data contained in each datachip. For a given region, only those
datachips whose bounding boxes overlap that of the region are extracted from the database
into memory. The spatial data in these datachips is then decoded and clipped while in
memory.
Datachip methods have been created for point, line and polygon data, and associated
sub-methods have been developed for performing the compression, decoding and other
manipulations. This new method is likely to become a vital tool in applications requiring
geospatial data processing and analysis.
|
datachips
