| Many geographical applications deal with objects in space that cannot be adequately described by determinate, crisp spatial concepts because of their intrinsically indeterminate and vague nature. Geographical information systems and spatial database systems are currently unable to cope with this kind of data. Large volumes of vague spatial data require a treatment in a database context for representing, querying, and manipulating them efficiently. To support such data and applications, we present a formal data model called Vague Spatial Algebra (VASA). This algebra comprises a set of vague spatial data types for vague points, vague lines, and vague regions together with a comprehensive collection of vague spatial operations and vague topological predicates. Besides its more fine-grained expressiveness than in the pure exact case, VASA's main benefit is that its formal framework is based on well known, general, and exact models of crisp spatial data types. This simplifies the definition of the vague spatial model since we can build upon an already existing theory of spatial data types. In particular, crisp spatial data types turn out to be a special case of their vague counterparts. In addition, our approach enables executable specifications for the operations, which can be immediately used as implementations. We present a prototype implementation of the algebra and of its underlying crisp spatial algebra in two dimensions (SPAL2D). We also provide the details of the mSLOB mechanism for handling multi-structured large objects in databases. The mechanism currently works as the generic interface between SPAL2D and any DBMS for which mSLOB can be implemented. My dissertation offers a precise and conceptually clean foundation for implementing a DBMS extension for vague spatial data and demonstrates the embedding of these new data types as attribute data types in a database schema as well as the incorporation of vague spatial operations and predicates into queries formulated in an SQL-like query language. |
