POINT REPRESENTATION OF SPATIAL OBJECTS AND QUERY WINDOW EXTENSION - A NEW TECHNIQUE FOR SPATIAL ACCESS METHODS

The use of Spatial Access Methods (SAMs) in spatial database systems, such as Geographical information Systems, is necessary to achieve effi cient retrieval of data items according to their spatial properties. E xisting SAMs organizing minimum bounding rectangle (MBR) approximation s of spatial objects can be classified into four groups. Each group is characterized by the special technique adopted for managing MBRs: (a) Ordering, (b) Transformation, (c) Clipping, and (d) Overlapping. This paper introduces a new technique. The basic idea of this technique is to represent all spatial objects by their MBRs and further reduce the m into points of the same dimensionality, so that any multidimensional Point Access Method (PAM) may be used to support access. Essential fo r the functionality of the new method is the query window extension. T he results of both analytical and experimental work show that SAMs usi ng the new technique clearly outperform popular SAMs, such as the R- a nd R-trees for data sets consisting of equal-sized MBRs. As for data sets of varying MBR sizes a competitive performance can be obtained.