Multi-Scale Spatial Database and Map Generalisation

The constitution of multi-scale spatial databases is closely linked to map generalisation. Elements of a multi-scale spatial database include multi-scale data models, multi-scale data storage schemes, methods to generate multi-scale datasets and methods to access these datasets and present results. In this paper we proposed a formal multi-scale spatial data model with an information-rich structure which supports representation of multiple objects at the same location and the same scale. In using existing generalisation procedures to generate multi-scale spatial datasets, an indicator which conforms to the retrieval precedence of vertices and a mapping mechanism to link scales and indicator values are required. The binary tolerance tree (BVT) is presented with the objective of meeting these two requirements. When data are retrieved from a multi-scale database, conflict resolution procedures can help to present the results in legible form. We present a shape-preserving method for on-line linear feature displacement using local coordinate transformation and analysis on curve characteristics. Finally, a brief introduction on several multi-scale geometry storage schemes as well as some conclusions drawn from experiments is given.