Identify Building Patterns

This paper presents a method of identifying patterns among buildings and constructing relationships between them. It is envisioned that the pattern objects not only provide operands upon which generalization can be applied, but more importantly, add a new spatial data type with specific semantics in a geospatial database. These new database objects can support spatial analysis and queries that make use of the semantics. In addition, spatial patterns at various scales conform to urban structures and form a natural hierarchy of map objects in a multi-resolution database, which facilitates a streamlined navigation and scalable visualization of cartographic maps. In this research, building patterns are identified based on user specifications which are accommodated by pattern templates. The parametric data model makes it possible to extend the library of pattern templates and to modify specifications to target intended spatial patterns. Map generalization with the knowledge of building patterns will be exemplified as an application of spatial patterns.