This paper focuses on how to flexibly access massive vector data on embedded devices. Combining hierarchical access spatial index and vector data simplification methods, a kind of vector pyramid model is put forward based on extensions of multi-scale R-tree. In the model, a sequence of coarse snapshots from a detailed map is created by vector compression methods according to the screen resolution， and the series of different scale snapshots is organized by a multi-scale R-tree of the source map and stored as a vector pyramid in advance. In this way, the embedded GIS applications can retrieve the proper details and features according to the display window. A test of forest sub-compartment data at the scale 1:10000 in Hunan is provided to show its efficiency.
展开▼
