ML-Quadtree: The Design of an Efficient Access Method for Spatial Database Systems

The aim of this paper is to present a new indexing technique that provides an efficient support for retrieving and handling spatial data. Traditionally, the mapping between layers (in a thematic point of view) and indexes structures is one to one. Each layer is associated with an index structure. In some previous work, we have presented a data structure, the FI-Quadtree that handles a set of images using only one index structure. This handling is a raster-oriented format. In this paper, we focus on the processing of these objects from the vector oriented format point of view. The Multi-Layer Quadtree (ML-Quadtree) is a new data structure that allows the storage and processing of several layers at the same time. This structure is based on the PM-Quadtree, which allows the storage of only a single-layer map. The aim of the MLQuadtree is to be able to manage, store and perform queries among multiple layers simultaneously. The design and the manipulation of the proposed structure is presented in this paper whereas the implementation and the experimentation result will be treated in a subsequent paper.