Spatial indexing in binary decomposition and spatial bounding

Abstract

In spatial database systems, data objects are of non-zero size and are associated with space coordinates. Efficient query retrieval on spatial relationships relies upon auxiliary data structures to support spatial indexing of these objects. The underlying indexing structures must support efficient spatial operations, such as locating the neighbors of an object. In Proc. IEEE Comput. Software & Applic. Conf, pp. 433–438 [1], we proposed an indexing structure called the skd-tree. This paper is a follow-up study of this initial design. In this paper, we describe the structure and related algorithms, the deletion algorithm in particular. Extensive experiments are presented. Performance study shows that the skd-tree is an efficient indexing mixture.